Betting

To Mum and Dad and my wife Julie

Professor Leighton Vaughan Williams

Betting To WinA Professional GuideTo Profitable Betting

HIGH STAKES

This edition published in May 2004 by

High Stakes Publishing

21 Great Ormond St

London WC1N 3JB

T: 020 7430 1021

www.highstakes.co.uk

www.gamblingbooks.co.uk/publishing

Copyright 2002 © Professor Leighton Vaughan Williams

The right of Professor Leighton Vaughan Williams to be identified as author of this work has

been asserted by him in accordance with the Copyright, Designs & Patents Act 1988.

All rights reserved. No part of this book may be reproduced, stored in or introduced into a

retrieval system, or transmitted, in any form or by any means (electronic, mechanical,

photocopying, recording or otherwise) without the written permission of the publishers.

Any person who does any unauthorised act in relation to this publication may be liable to

criminal prosecution and civil claims for damages.

ISBN 1-84344-015-6 BETTING TO WIN: A professional guide to

profitable betting

Typeset by Able Solutions (UK) Ltd, Birmingham

Printed by Cox & Wyman, Reading

Contents

Introduction 9

Part One: Betting With The Bookmaker 11

Chapter 1: Fair And Unfair Odds 13What’s Fair? ..................................................................................... 14The Way To Bet With The Bookmaker ............................................. 15Each Way Bets ................................................................................. 17Is Each Way Betting Good Value? ................................................... 18What Are Fair Place Odds? .............................................................. 20Multiple Bets .................................................................................... 22Forecast Bets ................................................................................... 24Choice .............................................................................................. 25Asian Handicap Betting .................................................................. 26

Chapter 2: Betting On The Tote 29The Way To Bet On The Tote ............................................................ 30

Chapter 3: Spread Betting 32The Way To Bet On The Spreads ..................................................... 35Risk Management In Simple Sports Spread Bets ........................... 36Financial Spread Betting ................................................................ 37Question .......................................................................................... 38Risk Management ........................................................................... 39Guaranteed Stop Loss ..................................................................... 39Managing Risk By Closing The Bet ................................................. 40Options ............................................................................................ 41

Chapter 4: Betting On The ‘Exchanges’ 44The Way To Bet On The Exchanges ................................................. 44

Chapter 5: Bookmakers And The Exchanges 49Comparing The Options .................................................................. 49

2.00 Cheltenham: Gerrard Supreme Novices’ Hurdle 492.35 Cheltenham: Irish Independent Arkle Challenge Trophy Chase 523.15 Smurfit Champion Hurdle Challenge Trophy 523.55 William Hill National Hunt Handicap Chase 544.30 Fulke Walwyn Kim Muir Challenge Cup Handicap Chase 55

5.05 Pertemps Final (Handicap Hurdle) 56

Punters Beware! .............................................................................. 57

Part Two: Beating The Bookmakers 61

Chapter 1: Theory And Evidence 63Chapter 2: How To Bet When The Odds Are In Your Favour 70Chapter 3: Should You Follow The Market? 72Chapter 4: Was The Steamer Value At The Off? 76Chapter 5: The Right Shape 79Chapter 6: The Gambler’s Fallacy 81Chapter 7: Do Multiple Bets Make Sense? 85Chapter 8: The Real Lesson Of Dettori Day 88Chapter 9: Draw Biases 91Chapter 10: What Goes Up, Must Come Down! 94Chapter 11: Betting In Running 96Chapter 12: The Difference Between Risk And Uncertainty 99Chapter 13: Strike While The Iron’s Hot 102Chapter 14: Prescribing The Right Dosage 104Chapter 15: The Home Ground Factor 106Chapter 16: The Patriot Bias 109Chapter 17: Small Is Beautiful 111Chapter 18: The Hi-Tech Crystal Ball 115Chapter 19: Better Late Than Ever! 117Chapter 20: The Value Of Tipping Agencies 119Chapter 21: Betting On The Oscars 122Chapter 22: Second Impressions Are Sometimes Better 126Chapter 23: The Problem With Miss World 127Chapter 24: Looking For Patterns 130Chapter 25: The Motivation Thing 132Chapter 26: Noise 135Chapter 27: Betting On A Leadership Election 137Chapter 28: You Can’t Beat Reliability 139Chapter 29: Cricket Betting 141

Part Three: Beating The Tote 143

Chapter 1: When Betting Longshots, Getting What You’re Given ReallyCan Be The Best Option 145Chapter 2: Steaming Into The Tote 147Chapter 3: Betting On The US Tote 149Chapter 4: US Horseracing On The TV 151

Part Four: Beating The Spreads 155

Chapter 1 : Profit Margins And Profitability In The Spread Markets 157Chapter 2: Avoid The Christmas Syndrome 159Chapter 3: Exploiting Disparities 161Chapter 4: To Arb Or Not To Arb: That Is The Question 163Chapter 5: Beware The Downside! 165Chapter 6: Quarb Strategy: Using The Odds To Beat The Odds 167Chapter 7: Tomorrow’s Another Day 170Chapter 8: Betting On The Election Spreads 172Chapter 9: Control The Consequences 174

Part Five: Beating The Exchanges 177

Chapter 1: Backing And Laying On The Exchanges 179Chapter 2: Odds Betting 180

Introduction .................................................................................. 180Chapter 3: Line And Range Betting 184

What Is Line Betting? .................................................................... 184What Is Range Betting? ................................................................. 185

Chapter 4: Asian Handicap Betting On The Exchanges 186Backing England At Scratch .......................................................... 187Backing England With A Half Goal Deficit .................................... 187

Part Six: World Cup Fever 189

Chapter 1: Different Betting Strategies 191Fixed Odds ..................................................................................... 191Spread Betting .............................................................................. 199Exchange Spread Betting ............................................................. 207

Chapter 2: If It Looks Too Good To Be True,Then It Is Too Good To Be True! 212

Conclusion 217

Appendix 1: Some Interesting And Useful Internet Sites 218Bookmakers .................................................................................. 218Spread Betting Companies ........................................................... 219Betting Exchanges ........................................................................ 219Sports And Betting Information And Advice ................................ 219

Appendix 2: World Cup 2002 Statistics 221Group A .......................................................................................... 221Group B .......................................................................................... 222Group C .......................................................................................... 223Group D ......................................................................................... 224Group E .......................................................................................... 225Group F .......................................................................................... 226Group G ......................................................................................... 227Group H ......................................................................................... 228Round of 16 ................................................................................... 229Quarter-finals ................................................................................ 231Semi-finals ..................................................................................... 2313rd place play-off .......................................................................... 232World Cup final .............................................................................. 232

Appendix 3:Poisson Table 233

Bibliography 234

Introduction

IF EVER THERE WAS A GOLDEN AGE of betting, this is it. Thechoice is extraordinary, the opportunities are bountiful, and at lastthe betting man or woman has a real chance of earning money atthe same time as enjoying the thrill of the chase.

In such an environment, there is a clear need for a book which isable to act as a guide through this choice, and through theopportunities on offer.

My research career, which has focused on the study of bettingand financial markets, pre-dates this Golden Age, to a time whenthe identification of profitable betting strategies was so much moredifficult. This served to ensure, of course, that the cutting edge ofthe analysis was sharpened all the more effectively. The tools thathave resulted are based on an extensive study of the most rigorouslyreviewed literature from across the world, and through original workof mine that has been published at the forefront of the internationalacademic literature. These tools are all the more valuable whenapplied to today’s punter-friendly betting environment.

I was approached to write this book to mark this Golden Age ofbetting. My remit was to provide a clear and accessible guide, basedon the most rigorous analysis available anywhere.

Betting has always been fun, but it has not always been profitable.Times have changed, and this book is my contribution to the newtimes. I hope you enjoy reading it as much as I have enjoyed writingit – and, most of all, I hope that it helps turn a pleasant hobby into aprofitable investment. Let battle commence!

Part One: Betting With The

Bookmaker

Chapter 1: Fair And Unfair Odds

THE ESSENTIAL IDEA behind bookmaking is a very simple one,and the example of setting odds about the outcome of a coin tosswill serve to explain it.

Clearly the outcome of a normal coin toss is restricted to twopossibilities, i.e. ‘heads’ or ‘tails’. If it is a fair coin, there is a 50 percent chance of heads and a 50 per cent chance of tails. This meansabout a half of all coins will end up heads and a half tails. Moreover,the more coins we toss, the closer we get to the even split predictedby theory.

If you are a bookmaker, and you are only interested in breakingeven, you will offer to match the punter ’s stake if he calls correctly,but keep his stake if he calls wrongly.

For example, if the punter calls ‘heads’, and stakes £10, you willtake the £10 and wait for the coin toss. If it lands on ‘heads’, thepunter wins, and so you match the £10 stake, and return the originalstake. This is known as giving odds of ‘even money’ or ‘evens’. If itlands on tails, however, the punter loses, and you keep the £10 stake.

This way, by offering ‘evens’, the bookmaker wins £10 every timethe punter is wrong, and loses £10 every time the punter is right.On average the punter is right half the time and wrong half thetime, so the bookmaker can expect to win £10 and lose £10 on anequal number of occasions.

The offer of ‘evens’ about a coin toss, or any other event with a50-50 chance of occurring, is known as an offer of ‘fair odds’.

Take now the case of dice. What is the chance that a single die, ifthrown, will show a six? Clearly the answer is 1 in 6. What is thechance that it will show a four? Again, the answer is 1 in 6. Thereason is that there are six possible outcomes, a 1,2,3,4,5 or 6, andthey all have an equal chance of coming up.

Before throwing the die, therefore, the fair odds about calling thecorrect number is 5 to 1. This means that the bookmaker should paythe punter 5 times his stake (as well as returning the stake) if the callis correct, but keep the stake if the call is wrong.

This way, the bookmaker wins £10 every time the punter is wrong,and loses £50 every time the punter is right. On average, the punter

14

BETTING TO WIN

is right one time in six, and so out of six throws, the bookmaker canexpect to win £10 on five occasions (a total of £50) and to lose £50 onone occasion. In other words, in the long run, the bookmaker canexpect to break even, as can the punter.

So much for fair odds! The bookmaker’s purpose, however, is tomake a profit, and that means offering less than fair odds.

Take again the case of the coin toss. The bookmaker who wishesto make a long-term profit might offer, say, 4 to 5 about the cointurning up ‘heads’ and 4 to 5 about it turning up ‘tails’.

This means that the bookmaker will pay out £4 if the punter callscorrectly, but keep £5 if the punter calls wrongly. Since the punterwill call correctly only half the time, on average the bookmaker wins£5 half the time, and loses £4 half the time. Clearly, at these odds,the bookmaker wins in the long run.

To summarise, if the odds set by the bookmaker are fair, thebookmaker and the punter will break even in the long run. If theodds are less than fair, the bookmaker will win, and the punter willlose, in the long run.

Now it would seem foolish to bet on the toss of a coin at less thanfair odds (evens), unless you enjoy the game, or unless you knowsomething you shouldn’t about the coin. The same applies to thedice game.

Of course, it is possible to win in the short run, even at less thanfair odds, by sheer luck. There is nothing, for example, to prevent acall of heads winning ten or twenty times in a row. This way ofthinking is not recommended, however, for those looking to a long-term betting strategy that will win money.

What’s Fair?In the case of the coin toss, or the dice game, it is easy (unless thegame is bent) to calculate what the fair odds should be. On thisbasis, it is straightforward to decide whether a bet at any given oddsis likely to yield a long-term profit, loss, or a break-even position.

What constitutes fair odds about a horse winning, or a dog, or afootball team, is altogether more difficult to identify, and two differentjudges may come up with totally different answers. It is thisdifference of opinion that has always allowed the betting public tolive in the hope that they can beat the bookie. The problem, of course,

15

FAIR AND UNFAIR ODDS

is that the bookmakers tend to get the odds right more often thanthe punter, and, besides, bookmakers are able to offer a set of oddswhich taken as a whole are designed in their favour. This margin intheir favour, or ‘over-round’, tends to be particularly large wherethe bettor has a relatively large number of alternatives to choosefrom, such as a 20-runner field in a horse race. In addition there hasuntil recently been the crippling effect of betting tax deductions.

The Golden Age of betting that we are presently experiencinghas brought with it changes in all these factors. First, the quantityand quality of information available to bettors is now so plentifulthat, if used properly, it can be used to significant effect in counteringthe advantage that the bookmaker has historically possessed in thesetting of odds, and particularly so where information is changingrapidly. Second, the number of different bookmakers is now soenormous that the bettor is able to shop around for odds in such away that at best odds the effective margin is wiped out, or eventurned to the advantage of the bettor. Finally, the switch from aturnover tax to a tax on the gross profits of bookmakers, and thesimultaneous reduction in the burden of taxation on bookmakers,has led to the present era of deduction-free betting.

The battle between bettor and bookmaker is at last set onsomething approaching a level playing field. Indeed, in many waysthe pitch is tilted to the advantage of the bettor. It is a battle thatreally can be won.

The Way To Bet With The BookmakerThe first thing to remember in the battle with the bookmaker is thatthere are many of them, but only one of you. Properly played, thatadvantage can prove decisive.

The first thing we shall do in this chapter, therefore, is to highlightthe number of bookmakers populating the market, and how bestvalue can be obtained by taking full advantage of the choice onoffer.

First, take note of the so-called ‘Big Three’, which has traditionallybeen made up of betting shops operated by Ladbrokes, William Hilland Coral, and arguably a ‘Big Four ’, which includes Totebookmakers.

These bookmakers currently operate in four distinct sectors: off-

16

BETTING TO WIN

course betting at licensed outlets (the biggest share of the market),on-course betting (betting at the racetrack), betting by telephone(through deposit or credit accounts, or via credit or debit cards),and betting via the Internet.

Bookmakers operate, especially on-course, by offering odds whichare fixed when the bet is struck. In this case, it doesn’t matter whenthe event takes place, the bet will be settled at the odds in placewhen the bet was agreed. If, for example, you bet on a horse at thetrack to win at 6 to 1, you will be paid out at this price regardless ofwhether the odds offered by the same or any other bookmaker getbigger or smaller. In other words, you are able to lock in a price.This is usually known as taking a price or taking the board price, thelatter term deriving from the historic practice of chalking up theprice on boards. This facility is also available off-course, in bettingshops, by telephone and on the Internet. A growing number ofbookmakers also offer their own set of early prices in the hours beforea race, or at least before the on-course market opens. The on-coursemarket for horse racing usually extends for about ten to twentyminutes before the start of the race.

Despite the availability of this option, most bets placed in bettingshops on horse or greyhound racing are struck at what is known asthe starting price (SP). The precise method of determining the startingprice has changed following recent reforms, but basically the startingprice represents the odds which independent assessors at theracetrack determine are generally available from bookmakers atthe start of a race.

The starting price about most horses is, on average, as long asit will ever have been during the on-course market. This is becausebookmakers are naturally very cautious when they display oddsfor the first time. It is only when they are able to gauge the relativesupport for different horses that they will start to offer betterodds so as to attract customers. The starting price will also usuallybe longer than the typical ‘early price’ offered by any givenbookmaker about a horse.

There is one exception, however, and it is an important one.Horses that go on to win tend to attract significant support fromclued-up bettors before the start of the race. For this reason, thestarting price about these horses is often shorter than prices that

17


could have been taken earlier. The problem, of course, is thatunless you know in advance which horse is going to win, youwill lose out in the long run by grabbing the earlier odds.

Finally, there is no convincing evidence that the last board priceavailable before the start of the race is significantly better or worsethan the officially returned starting price.

In addition to board prices or early prices, most bookmakersalso offer ante-post prices, which are odds laid before the day ofthe race or the event. These prices will often be more generousthan the subsequent early price about the same selection or thesequence of board prices or the starting price. There is a criticaldisadvantage in taking an ante-post price, however, in that if yourselection withdraws, for whatever reason, after the bet is struck,you will in most circumstances lose your stake.

If you take an early price or a board price on a horse, however,and it subsequently withdraws (technically, does not come understarter ’s orders), your stake is returned. This sometimes happensafter an early price has been taken, or in the case of the startingprice so late that bookmakers are unable to compile a fresh set ofodds to allow for this. To compensate, bookmakers deduct someof the payout to the winning bets. This so-called ‘Rule 4’deduction is calculated on the basis of the odds prevailing aboutthe withdrawn horse at the time it was withdrawn.

In addition to straight win bets, there are also a wide range ofother bets available, of which ‘each way’ bets lead the way. Theseallow the bettor to nominate a horse either to win or be placed(usually, but not always, in the first three). There are also multiplebets on cumulative outcomes and forecast bets such as theComputer Straight Forecast and Tricast, which involvenomination of the first two or three past the post in the correctorder.

Each Way BetsIn each way betting, you bet on your selection to win or to beplaced. For example, instead of betting £20 to win, you bet £10each way, i.e. £10 to win and £10 to be placed.

The definition of a place differs depending on the event andthe number of entrants, but usually it means to be placed in the

18

BETTING TO WIN

first three. In horse racing, this is strictly defined. In a race of five, sixor seven runners, a place means finishing in the first two; in races ofeight or more runners it usually means to finish in the first three, orthe first four in handicaps of sixteen or more runners. Sometimebookmakers offer more generous terms, including a place for the firstfive home in some high-profile handicaps. For the record, handicapraces are races where the horses carry different weights, dependingon their ability, in order to make the race more competitive.

The usual place terms are shown below:

5 to 7 runners: 1/4 odds – 1st or 2nd – all races

8 or more: 1/5 odds – 1st, 2nd or 3rd – all non-handicaps

8 to 11 runners: 1/5 odds – 1st, 2nd or 3rd – all races

12 to 15 runners: 1/4 odds – 1st, 2nd or 3rd – handicaps only

16 or more runners: 1/4 odds – 1st, 2nd, 3rd or 4th – handicaps only

It is also possible to make a win and place bet, which meansstaking to win a greater amount than to place. Bookmakers willnot normally accept a bet just to place, however, or to place for agreater stake than to win.

Is Each Way Betting Good Value?The simplest way to consider the relative value of a win bet comparedto an each way bet is to take a simple example.

The example is of a six-horse race, in which all the horses arejudged to be equally good, and the bookmaker takes no margin.This is an unlikely case, of course, but it serves to illustrate theprinciple very well, and the same logic applies to cases of unequallymatched horses, and/or where the bookmaker builds a fat margininto the odds.

In this mythical six-horse race, the fair bookie offers odds of 5 to 1about each equally matched horse. You could in this case place £20on each of the six horses to win, at 5 to 1, amounting to a total stakeof £120.

Whichever horse wins, your return would be 5 x £20 (since youhave staked £20 at 5 to 1), plus your stake returned, i.e. £120.

You and the bookmaker break even.

19


Assume now that instead of betting £20 on each horse, you bet£10 each way on each (i.e. a total of £20 per horse). This is made upof £10 to win and £10 to place on each horse. Again your total stakeis £120.

However, only one horse will win, say Horse A, and another willfinish second, say Horse B.

The return to the winning horse is calculated as follows:

£10 to win on Horse A at 5 to 1 = 5 × £10 plus £10 stake returned = £60.

A winning horse is also considered to be placed (1st is a place aswell as a win).

The £10 to place on Horse A yields 1/4 of the odds of 5 to 1, plusthe stake returned, i.e.

1/4 x 5 × £10 plus £10 stake returned = £12.50 + £10 = £22.50.

The £10 to win on Horse B is clearly lost.However, the £10 to place on Horse B yields the same as for Horse

A, i.e. £22.50 (since the odds are the same).So the total return to your total stake of £120 (£60 each way) is:

£60 + £22.50 + £22.50 = £105.

You have, therefore, staked £120 to ensure a return of £105.

To summarise

£20 to win at 5 to 1 on six horses costs £120.Return whichever horse wins = £120.£10 each way at 5 to 1 on six horses costs £120.At place odds of a quarter of the win odds, a £10 each way bet on

each horse yields the following:Total cost = 6 × £20 = £120.Return on winning horse = win return + place return.Win return = £10 at 5 to 1 plus stake returned = £60.Place return = £10 at 0.25 (5 to 1) plus stake returned = (1.25 ×

£10) + £10 = £22.50.

20

BETTING TO WIN

Return on second-placed horse = win return + place return.Win return = 0.Place return = £10 at 0.25 (5 to 1) plus stake returned = (1.25 ×

£10) + £10 = £22.50.Total return to each way backers = £105.Compared to the win bet, where you broke even, this time you

have lost £15.

What Are Fair Place Odds?For the win bet and the each way bet to represent equal value, theplace odds would in this example need to have been about two-fifths rather than a quarter of the win odds.

This can be demonstrated as follows:£20 to win at 5 to 1 on six horses costs £120.Return whichever horse wins = £120.£10 each way at 5 to 1 on six horses costs £120.At place odds of 0.4 of the win odds, a £10 each way bet on each

horse yields the following:Total cost = 6 × £20 = £120.Return on winning horse = win return + place return.Win return = £10 at 5 to 1 plus stake returned = £60.Place return = £10 at 0.4 (5 to 1) plus stake returned = (2.0 × £10)

+ £10 = £30.Return on second-placed horse = win return + place return.Win return = 0.Place return = £10 at 0.4 (5 to 1) plus stake returned = (2.0 × £10)

+ £10 = £30.Total return to each way backers = £120.In other words, for each way odds to represent the same value as

win odds, the place part of the bet should be calculated as 0.4 timesthe win odds.

In actual fact, the place odds in this example are calculated as aquarter (0.25) times the win odds.

In the case of six-horse races at current place odds, therefore, winbets would seem to represent better value than their each wayequivalents.

For each level of win odds, the following table shows what fractionof the win odds the place odds should be in a fair book, and what

21


fraction they actually are. Win bets only are accepted if there arefewer than five runners.

HandicapsRunners Fair place fraction Actual place fraction No. placed

5 0.38 0.25 2

6 0.40 0.25 2

7 0.42 0.25 2

8 0.24 0.20 3

10 0.26 0.20 3

12 0.27 0.25 3

14 0.28 0.25 3

16 0.20 0.25 4

20 0.21 0.25 4

Non-handicapsRunners Fair place fraction Actual place fraction No. placed

5 0.38 0.25 2

6 0.40 0.25 2

7 0.42 0.25 2

8 0.24 0.20 3

10 0.26 0.20 3

12 0.27 0.20 3

14 0.28 0.20 3

16 0.29 0.20 3

20 0.30 0.20 3

According to this analysis, place odds (and therefore each waybets) represent worse value than a straight win bet in all cases excepthandicap races of 16 or more runners.

Even so, there may be occasional cases where you have goodreason for believing that a horse is disproportionately likely to beplaced, but is very unlikely to win, and that the each way bet maytherefore represent value. These are exceptions to the general rule,

22

BETTING TO WIN

however, and should in each case be carefully considered on theirindividual merits.

Multiple BetsThere exist a range of multiple bets, i.e. bets that yield a return ifmore than one part of the bet comes up. The simplest is a win double.In a win double, the bettor chooses two selections. To win, both ofthe selections must be successful. If both horses win, the odds aremultiplied. If either loses, the bet is lost.

ExampleYou select Horse A to win the first race on the card at 2 to 1. You

select Horse B to win the second race on the card, at 10 to 1. Younow ask for a £10 win double on these two horses.

The double is calculated like this. Your £10 stake goes first on toHorse A. If it wins, your return is £30 (2 times £10 plus your stakereturned). That £30 is then placed on Horse B. If Horse B wins, yourreturn is £330 (£30 at 10 to 1 yields £300, plus your £30 stake returned).

All multiple bets follow this principle.A common alternative to the win double is the win treble. These

operate exactly like doubles, except that there are three selectionsinstead of two. Accumulators can be constructed in the same manner,built up of four or more selections.

To calculate your return, convert the odds on offer into decimalnotation, if they are not shown that way to start with. To do this,add one to the traditional (or fractional) odds. For instance, odds of2 to 1 are represented in decimal notation as odds of 3.0. Odds of 7to 2 (3.5 to 1) are represented, in decimal notation, as odds of 4.5.

Now multiply these decimal odds together.So, if you have three selections at 2 to 1, 3 to 1 and 4 to 1, these are

converted into their decimal equivalents of 3.0, 4.0 and 5.0. If theyall win the return is calculated by multiplying these odds together,i.e. 3 x 4 x 5 = 60.0.

On this basis, a £20 win treble, for example, at the odds shownabove would yield a return of 60 × £20.00, i.e. £1200.

There are numerous multiple bets, usually with strange names,which simply combine a number of different multiple bets. Forexample, if you make five selections, these can be combined into

23


doubles, trebles, four-selection accumulators, and five-selectionaccumulators. This is obviously rather more costly for a given unitstake, but the advantage is that you will get some return if onlysome of your selections win. Bookmakers often offer incentives topunters to make these sorts of wagers, usually by offering a bonuspayout to winning bets.

Some common examples of this type of multiple bets are as follows:

Trixie three selections, four bets (three doubles, one treble)

Patent three selections, seven bets (three singles, three doubles,

one treble)

Yankee four selections, eleven bets (six doubles, four trebles, one

fourfold)

Lucky 15 four selections, 15 bets (four singles, six doubles, four trebles,

one fourfold).

Canadian five selections, 26 bets (ten doubles, ten trebles, five

fourfolds, one fivefold)

Lucky 31 five selections, 31 bets (five singles, ten doubles, ten trebles,

five fourfolds, one fivefold)

Heinz six selections, 57 bets (fifteen doubles, twenty trebles, fifteen

fourfolds, six fivefolds, one sevenfold)

Lucky 63 six selections, 63 bets (six singles, fifteen doubles, twenty

trebles, fifteen fourfolds, six fivefolds, one sevenfold).

Super Heinz seven selections, 120 bets (21 doubles, 35 trebles, 35

fourfolds, 21 fivefolds, 7 sixfolds and one sevenfold).

Goliath eight selections, 247 bets, 28 doubles, 56 trebles, 70 fourfolds,

56 fivefolds, 28 sixfolds, eight sevenfolds, one eightfold.

In addition to win multiples, bettors can also stake each waymultiples. Each part of the selection is then treated as an each waybet. Since the each way bet is made up of a bet to win and a bet toplace, naturally the number of bets is doubled when placing theeach way equivalent of each of the above.

When betting tax was levied on the stake, the advantage of themultiple bet option was that the same tax was levied on a multipleof a given stake, say £10, as on a single of the same stake. The potentialwinnings were higher, therefore, for the same tax.

Since the abolition of betting deductions, there is no advantage

24

BETTING TO WIN

to be gained from placing a double on two horses over placing twosingle bets, except perhaps in terms of time saved and convenience.To convert two single bets into a double, one need simply bet on thefirst horse. If it wins, stake all the winnings on the second horse.This is exactly what the double does automatically. However, byplacing two singles you are able to control when and how muchyou bet on the second horse.

The same logic applies to a treble (all three selections have to besuccessful) or indeed any multiple bet.

Sometimes, however, the bookmaker will try to encourage multiplebets, by adding a bonus to any that win. The bookmaker’s reasoningis that any advantage over the punter is multiplied in multiple bets,and so these bets tend to be especially profitable from thebookmaker ’s point of view. Unless you are confident that theadvantage is actually in your favour at each stage of the bet, therefore,think very hard before placing a multiple. After all, the great flexibilityof choosing when and how much to bet, and at what odds, isdiminished once you lock your money in early.

Forecast BetsForecast bets are similar to multiple bets, except that the bet involvesmore than one horse in the same race rather than different races.

The best-known example of a forecast bet in the UK is thebookmakers’ Computer Straight Forecast. This involves selection ofthe first two horses to pass the post in the correct order.Unfortunately, there is no simple way of working out in advancehow much you will win if you are right. Instead, the payout iscalculated according to a complex formula devised by thebookmakers. Need one say more?

The Tote’s competitor to the Computer Straight Forecast used totake the form of the Dual Forecast. In the Dual Forecast punterswould specify the first two home in either order (rather than thecorrect order), and the pool of winning bets was shared (after a 24%deduction).

Most independent analysis suggested that the Dual Forecast offeredbetter value, in most circumstances, than the Computer StraightForecast. The Dual Forecast has now been replaced by the Tote Exacta.This is the Tote’s direct competitor to the Computer Straight Forecast.

25


If you must do forecast bets, it is worth checking a sample ofpayouts from the Computer Straight Forecast and the Exacta toidentify for yourself which offers the better value. Most studiessuggest that the answer is the Exacta, despite the sizeable deductionfrom the pool.

Other common forecast bets are the bookmakers’ Tricast (first threein the correct order) and the Tote Trifecta (a pool bet with returnsshared among those selecting the first three in the correct order).

Forecast bets are not available on every race. It is important tocheck with the individual rules of the operator before assuming thata Forecast bet is applicable.

ChoiceThe rise of telephone betting, and latterly betting on the Internet,has led to an explosion of choice for those seeking value.

Perhaps the best-known forum for highlighting the abundanceof choice, and for pinpointing value, is the Racing Post’s ‘Pricewise’column. This has been hosted by a succession of tipsters, includingmost recently Henry Rix and Melvyn Collier, both of whom havegone on to tip on a private subscription basis. At the time of writingthe incumbent Pricewise is Tom Segal.

The Pricewise column, which appears every Saturday, andperiodically on other days of the week, highlights a selection ofbookmakers who are offering early prices about each of the horsescontesting selected races. These races are usually, though not always,broadcast live on terrestrial TV.

Pricewise by no means highlights all the offers available, but it isa good start, and it is complemented by advice and analysis.

Latterly, those seeking to compare the odds on offer for best valuehave had the benefit of a number of Internet sites which have beenset up to cater for this very need.

A good example is Oddschecker. Oddschecker is available atwww.oddschecker.com and at www.oddschecker.co.uk

Easyodds is another odds comparison service, available atwww.easyodds.com

Bestbetting can be found at www.bestbetting.comOther useful odds comparison sites include:

26

BETTING TO WIN

• www.smartbet.com• www.bookiesindex.com• www.betbrain.com• www.crastinum.com• www.bookiebusters.net

Asian Handicap BettingAsian handicap betting is an increasingly popular form of betting,particularly for match bets where one side of the match bet issignificantly superior to the other. The basis of Asian handicapbetting lies in the way in which it attempts to even up the chancesof both sides of the match.

To achieve this evening up process, the side judged to be superioris artificially handicapped for betting purposes by the market-maker,relative to the side judged to be inferior. In other words, the sideperceived to be the weaker of the two is awarded a notional headstart as far as the match betting is concerned. The consequence isthat the odds about the two options are brought closer to each otherthan would be the case if no handicapping process were involved.

Asian handicap betting has two forms.The first is the single handicap, and the second is the dual handicap.

In the case of the latter, you are effectively placing two bets, each ofwhich is worth a half of your total stake.

Consider first the basic single handicap, and take for the sake ofsimplicity a simple half goal handicap in favour of the away side.

Example: Man Utd (-½) v. Man City (+ ½)In this example, Manchester City, playing away, are given an

artificial half goal start over Manchester United for betting purposes.With the benefit of this advantage, if Man City win or draw thematch, they (Man City) will be declared the winners for bettingpurposes. If Man Utd win the match, then all bets placed on ManCity will be losers. This is because Man Utd will have scored at leastone goal more than their opponents, which is sufficient to outweighthe notional half-goal advantage handed to the visitors.

Assume now that the perceived margin of superiority is greaterthan this.

27


Example: Arsenal (-1½) v. West Ham (+ 1½)Here West Ham are given a 1½ goal start over Arsenal for betting

purposes. If Arsenal win the game by at least two goals, all betsplaced on Arsenal will be winning bets. If West Ham win, draw orlose by a margin of just one goal, however, then all bets on WestHam are judged to be winning bets.

Take the case where both teams are judged by the market-makersto be evenly matched. This is shown on the handicap as follows:

Example: Tottenham (level) v. Leeds (level)In these circumstances, stakes are simply returned if the match is

drawn. If Tottenham win, on the other hand, all bets on Tottenhamare winning bets. Similarly, if Leeds win, all bets on Leeds arewinning bets.

Take now the case of a handicap of exactly one goal.

Example: Newcastle (-1) v. Sunderland (+ 1)In this case, Sunderland are given a one goal start over Newcastle.

In these circumstances, if Sunderland win or if the game is drawn,then bets placed on Sunderland will be winning bets. If Newcastlewin by at least two goals, bets on Newcastle will be winning bets. IfNewcastle win by one goal, however, the handicap bet is even andso all stakes are returned.

As long as the handicap line includes a half goal, a handicap drawis of course not possible.

In those cases where the handicap includes a whole number, ahandicap draw is a possible outcome, in which case all stakes arereturned.

There are more complex outcomes, where two handicaps areinvolved. In these circumstances, half your stake is placed at onehandicap and half at the other.

Example: Liverpool (-0, -½) v. Chelsea (+ 0, + ½)This handicap actually represents two bets – half your stake at

the 0, or level, handicap and half your stake at the half goal handicapin favour of Chelsea.

In this case, if you bet on Chelsea and the game is drawn, one

28

BETTING TO WIN

half of your bet is a winner (the handicap bet at + ½) and the otherhalf of your bet is effectively void (the bet at level handicap) andthat part of your stake is returned. If you bet on Liverpool, and thegame is drawn, again half of the bet is effectively void (that portionof the stake is returned) and the remaining half of the total stake (onthe handicap bet at - ½) is lost.

Example: Aston Villa (-½, -1) v. Southampton (+ ½, + 1)In this case, your total stake is divided equally between the ½ and

1 goal handicap bets in favour of Southampton. In thesecircumstances if you bet on Southampton, and Aston Villa beatSouthampton by a single goal, one half of your bet is cancelled outby the one goal handicap, and therefore one half of your total stakeis returned. The other half of your bet is a losing bet, since the halfgoal handicap does not cover the one goal deficit. If you bet onAston Villa, and Villa win by a single goal, again half the bet iseffectively void, and this portion of the stake will be returned onthe - 1 goal handicap. On the other hand, half of the bet will be awinning bet, since the one goal victory is not cancelled out by the ½goal handicap advantage enjoyed by Southampton.

29

BETTING ON THE TOTE

Chapter 2: Betting On The Tote

In the UK there is an alternative to betting with bookmakers, andthat is betting with the Tote (the Horserace Totalisator Board). TheTote is a pool form of betting, whereby all money staked is pooledtogether, and the pot (after deductions) is shared out among winningpunters.

To the operators of the pool, the exercise is risk-free, as the moneypaid out is simply a share of what is already paid in. Since there aredeductions taken from the pool, punters as a whole earn less thanthey bet, and so the average bettor loses. This is not to say that askilled bettor cannot win in the long run by betting with the Tote,but it does mean that the bettor has to be able to overcomedeductions of up to 30%.

There are a number of different pools, and the lowest deductionis from the win pool, i.e. the pool of all bets staked on a horse (ordog) to win. The largest deduction (of 30%) is from the so-called‘Scoop 6’ pool, which is a pool of bets seeking to identify the winnersof six selected televised races. The chances of winning are, of course,very small, but the winner or winners are assured a commensuratelysizeable payout.

The real disadvantage of betting on the Tote is that you do nothave the option of learning what odds you are taking at the timethe bet is struck. The final odds depend on the relative amounts ofmoney staked on the different outcomes, which is not known forcertain until after the race commences. Even so, an indication isavailable from computerised screens that provide regularly updatedinformation on flows of money into the pool. A related problem isthat one very big bet can change the odds sharply, particularly atmeetings where the pool is relatively small. High rollers are in thesecircumstances advised to stay clear, for their own stake may wellcut the payout about their chosen beast quite drastically.

On the plus side, the Tote is totally impartial, and so the odds onoffer genuinely reflect the weight of money placed by punters. Inparticular, horses (or dogs) which are unpopular with the bettingpublic will start at good odds, i.e. they will pay handsomely to thosefew who do back them, if they win. There is ample published

30

BETTING TO WIN

evidence which shows a tendency for the odds available about suchso-called longshots with the Tote to be rather more generous thanwith the bookmakers. However one sees it, the choice that the twobetting mediums offer the UK punter is clearly of significant benefit,if used properly.

Most countries do not offer bettors the luxury of betting withbookmakers or with the Tote. In the USA, for example, horse racingand greyhound racing are operated by a Tote (so-called parimutuel)monopoly. Bookmakers at the track are outlawed. The same appliesin many other countries, such as France, Canada, Hong Kong, Japanand New Zealand.

Australia has an in-between system, whereby the Tote andbookmakers co-exist at the racetrack, but where the Tote has amonopoly off-course.

To summarise, the co-existence of bookmakers and the Tote offersa clear benefit to the betting public, who can choose the better oddson offer. For less fancied runners, in particular, this is often availablewith the Tote. Either way, it is worth comparing the odds beforeplacing a bet, while taking account of the fact that it is not possibleto know with certainty what odds you are getting unless you take aprice with a bookmaker.

The Way To Bet On The ToteThe Tote operates as both a bookmaker and as a monopoly supplierof pool betting on horse and dog racing. It was set up as a poolbetting operation in 1929, but this now makes up only a smallproportion of its total turnover.

From the astute bettor ’s point of view, this is actually a littlesurprising, because the odds available about the same horse orgreyhound in the Tote pool is often superior to that available withbookmakers at starting price, and on the average significantly betterfor less-fancied chances.

The basic way of betting on the Tote at the course is to approachany of the red-coated attendants manning the Tote terminals whichare usually conveniently near to the paddock or hospitality areas ofthe racecourse. Tote Credit clients also have access to their own clientarea. Off-course, bets can be placed at any Tote betting office, or by

31

BETTING ON THE TOTE

telephone, or on the Tote web-site. There are also a number of so-called Tote Direct terminals in the offices of rival bookmakers, notablyLadbrokes and Coral, which accept bets straight into the pool.

The simplest bet is a bet to win. The minimum stake is £2, and thewhole of the pool is shared out among winning bettors, save for adeduction of 13.5%. There is no maximum stake but in a small poola large bet is likely to depress the odds about the selection too muchto make betting sense. In the larger pools, however, characteristic ofall-Tote countries, especially Hong Kong, even sizeable bets may failto make a dent in the odds.

Other Tote bets include:

Place: this is a bet on the horse to finish in the first three (usually),

although the number of finishers which qualify for a

winning place bet can vary depending on the number of

runners and the type of race.

Each way: this is a win and a place bet on the same horse. It is

available on all races of five or more runners.

Exacta: this bet requires selection of the first and second-placed

horse in the correct order.

Jackpot: this bet requires selection of the winners of all six specified

Jackpot races (usually races 1 to 6). It operates daily at a

selected meeting.

Placepot: This bet requires selection of a placed horse in all of six

specified Placepot races (usually races 1 to 6). It operates

at every meeting.

Quadpot: This bet requires selection of a placed horse in each of four

selected Quadpot races (usually races 3 to 6). It operates at

almost every meeting.

Trifecta: This requires selection of 1st, 2nd and 3rd in the correct order

in nominated races of eight or more runners.

Scoop 6: This requires the selection of the winners of each of six

nominated TV races, and operates on a Saturday only. The

races may be selected from more than one meeting.

32

BETTING TO WIN

Chapter 3: Spread Betting

In addition to the well-established Tote (parimutuel) system ofbetting, and odds offered by bookmakers, a third method of bettingon horse racing and indeed a wide array of other events hasdeveloped in recent years, called spread (or index) betting.

In spread betting in a sports context, the event that is the subjectof the bet may be almost anything. Popular examples include thenumber of goals (or indeed bookings) in a football match, the numberof runs scored by a team (or individual player) in a cricket match,the winning distance in a horse race, the number of points scoredin a rugby match, and so on.

For financial spread betting, the commodity is usually the priceof a stock, or the level of an index, at some specific time in the future,say the close of the present trading day, or a given time on a givenday in a given month. Spread betting enables a trader to take aposition on the price of a stock without any need ever to trade thestock itself.

Whatever the context, the market-makers set a ‘spread’ aroundthe expected outcome, and their clients (the bettors) are invited tobuy at the top end of the spread, or sell at the bottom end.

For example, in a cricket match between England and Australia,the spread betting company’s odds-setter might estimate thatEngland will score 200 runs in their first innings. In this case, a‘spread’ might be set of, say, 190 to 210 runs, which includes theestimate. In this example 190 runs (the lower figure) is known asthe bottom end of the spread, while 210 runs (the higher figure) isknown as the top end of the spread.

A bettor who believes that England will perform rather worsethan expected may now sell at the bottom end of the spread, at 190runs, for any given stake, within his credit limit, say £10 a run. IfEngland were to score less than 190 runs, the bettor would win thestake (£10) multiplied by the number of runs by which England fallshort of 190. For example, if England scored 170 runs, the bettorwins £200, calculated as the difference between the final outcome(170) and the sell point (190) multiplied by the stake (£10). On theother hand if England were to score more than 190, the bettor would

33

SPREAD BETTING

lose the stake (£10) multiplied by the number of runs by whichEngland exceed 190. For example, if England score 270 runs, thebettor loses £800, calculated as the difference between the finaloutcome (270) and the sell point (190) multiplied by the stake (£10).

The same reasoning applies to a buy bet. In the example of theEngland – Australia game, the spread on England’s first innings runstotal is set at 190 – 210. A bettor who believes that England will performrather better than expected may buy at the top end of the spread, at210 runs, for any given stake, within his credit limit, say £10 a run. IfEngland were now to score more than 210 runs, the bettor wouldwin the stake (£10) multiplied by the number of runs by whichEngland exceed 210. For example, if England score 270 runs, the bettorwins £600, calculated as the difference between the final outcome(270) and the buy point (210) multiplied by the stake (£10). On theother hand, if England were to score less than 210, the bettor wouldlose the stake (£10) multiplied by the number of runs by whichEngland fall short of 210. For example, if England score 170 runs, thebettor loses £400, calculated as the difference between the finaloutcome (170) and the buy point (210) multiplied by the stake (£10).

To take another example, we can consider the market about thetime that the first goal will be scored in a football match.

Say, for example, the bookmaker’s best estimate of the time thatthe first goal will be scored is 35 minutes. In this case, the spreadmay be quoted as, say, 34 – 36. The bettor may now buy the time ofthe first goal at 36 or sell at 34, for a given stake. If the bettor buys at36 and the first goal is scored in, say, 44 minutes, the bettor wins 8(the difference between 44 and 36) times the stake. If, on the otherhand, the first goal was scored in 24 minutes, the bettor would lose12 (the difference between 24 and 36) times the stake. The samelogic applies to a sell, so that the bettor wins (loses) in proportion tohow right (wrong) the original bet turned out to be.

There are currently four major sports spread betting companiesfor sport. These are Sporting Index, IG Index, Cantor Sport andSpreadex. City Index specialises in spread betting on financialmatters, such as the price of shares at the close of trading on thestock market. All of these companies are regulated by the FinancialServices Authority, and each company may offer a different quoteabout the same market.

34

BETTING TO WIN

Another relatively new player on the market is Sportsspread.Financial spread betting is offered by Cantor Index, IG Index, City

Index, Spreadex, Finspreads, Tradindex and Deal4free.Whether in sport or in the stock market, spread betting profits

are free from capital gains tax, and there are no dealing orcommission charges, stamp duty or other deductions. Oneconsequence of the regulatory regime is that, unlike bets withUK bookmakers which are binding in honour only, a spreadbetting transaction is a legally enforceable contract; bettors mayincur heavy losses if events turn against them.

A gambler wishing to participate in spread betting must normallyarrange a suitable amount of credit with a spread betting firm, andspecific transactions are usually conducted by telephone, with thecalls recorded to settle any dispute. The firm may decline any bet agambler proposes, or accept it only at a lower unit stake, or else mayask the gambler to increase his line of credit before the bet can beaccepted. Clients may open or close trades at any time in dealinghours, with instant execution.

A common use of this flexibility when trading in single shares isto make a bet that a price will fall, or rise. For example, if the buyprice of a share is 500p, a gambler who expected the price to risecould decide to buy at the rate of £10 for each 1p movement in theprice. The maximum loss, if the share became worthless, would be500 x £10, i.e. £5000.

The scope of ‘assets’ that are traded in spread betting marketsis wide, ranging from the price of gold to the number of goalsin a soccer match. Indeed, the ‘asset’ that is the subject of thetrade can include almost any clearly quantifiable feature of anarray of sports, political and financial markets. There are alsosome so-called ‘speciality ’ indices, for example the number ofOscars won by a given film at the Academy Awards ceremony.Standard political trades include the number of seats gainedby a poli t ical party in an elect ion. Financial trading isexceptionally diverse, and includes the values of the WallStreet, DAX, FTSE, Hang Seng and Nikkei share indices, theprice of individual shares traded on these markets, the priceof a variety of commodities, as well as bond and currencyfutures.

35

SPREAD BETTING

The clear advantage of spread betting, therefore, is the sheerdiversity of eventualities on which a client can bet. Moreover, thebettor can win large sums of money for a relatively small unit stake.Therein, however, lies the major disadvantage of spread betting.The sums lost can exceed the sum staked by a huge amount, unlikefixed odds or pool betting, where the most that can be lost is thesum staked.

In summary, spread betting is a relatively novel form of betting,which offers the opportunity to win (or lose) large amounts ofmoney. In the right hands, it can be an invaluable part of the bettor ’sarmoury, but because losses are not limited to the stake, extremecaution should be exercised, particularly by the novice punter.

The Way To Bet On The Spreads

Test YourselfYour first bet

You wish to bet on the number of corners in a forthcoming matchbetween Leeds and Liverpool.

You are offered a quote of 10 – 11.You think there will be more than 11 corners, and you buy at 11

for £20 per corner.

IF FINAL RESULT = 15 corners, what is your profit/loss?

AnswerProfit = Stake (Outcome - Buy price)

= £20 (15 - 11) equals £20 (4)= £80

IF FINAL RESULT = 9 corners, what is your profit/loss?

AnswerProfit = Stake x (Outcome - Buy price)

= £20 (9 - 11)= - £40

LOSS = £40

36

BETTING TO WIN

Your Second BetYou are offered a quote of 10 – 11.You think there will be fewer than 10 corners, and you sell at 10

for £15 per corner.

IF FINAL RESULT = 8 corners, what is your profit/loss?Answer

Profit = Stake x (Sell price - Outcome)= £15 (10 - 8)= £30

IF FINAL RESULT = 18 corners.Answer

Profit = Stake x (Sell price - Outcome)= £15 (10 - 18)= - £120

LOSS = £120

Risk Management In Simple Sports Spread BetsIn each of these cases you can close the bet at any time during theopening of the market, by making a bet in the opposite direction ofthe same stake.If this is at the same quote as that at which you opened the bet, youwill lose your stake multiplied by the size of the spread.

ExampleAt the start of the match, the quote about the number of corners

is 10 – 11.You buy corners at 11, for £20 per corner.If you now change your mind and decide to close the bet entirely,

you must sell at 10 for £20.You lose: (buy price - sell price) stake

= (11 - 10) £20= £20

The usefulness of a strategy of closing the bet often occurs whenthe game is in play. A number of markets, particularly live televisedmatches, are traded ‘in running’.

37

SPREAD BETTING

Take the same example as above: You buy corners at 11 before thematch.

Now examine the following scenarios.

Scenario 1After ten minutes of play, four corners have been taken.The quote has risen to 13 – 14.If you wish to close the bet totally, you can sell at 13, for the same

stake (£20).You now earn a sure profit of (new sell price - old buy price)

stake = (13 - 11) £20 = £40.

Scenario 2After 30 minutes of play, no corners have been taken.The quote has now fallen to 7 – 8.You decide to close the bet, by a sell at 7 for £20.You incur a sure loss of (old buy price - new sell price) stake =

(11 - 7) £20 = £80.In each case, you have decided to close the bet to guarantee a

given profit or to cap a given loss.

Financial Spread BettingAn increasingly common way of trading, particularly suitable forthe smaller investor, is known as financial spread (or index) betting.

Financial spread betting companies operate, just like sports spreadbetting companies, by quoting a spread about variables in a marketcharacterised by an uncertain future outcome. A popular exampleis the FTSE 100 index, i.e. the index of the performance of the majorstocks quoted in the UK. The higher the FTSE 100 index, the higherthe price, on average, at which the top shares are trading.

Say, for example, that the market-makers consider that the FTSE100 will close the day at 5000.

In this case they may quote a spread of, say, 4995 – 5005, whichincludes their best estimate of the actual closing price.

Clients of these companies are now invited to buy at the top endof the spread (5005) or sell at the bottom end (4995). The outcome ofthe trade is calculated as the number of units by which the actual

38

BETTING TO WIN

outcome differs from the level at which the trade is enacted.The units are calculated according to the minimum marginal

difference between positions at which trades can be implemented.The profit/loss is calculated as this number multiplied by the unitstake at which the client places the trade.

ExampleOpening of market: FTSE 100 is trading at 5015The market-makers quote a spread on the day’s closing price of,

say, 5010 – 5020One hour after the opening of the market: FTSE 100 has risen to

5025The market-makers are now offering a spread of 5020 – 5030.You have two choices if you wish to trade.Choice A: buy at 5030.Choice B: sell at 5020.You must then choose your unit stake, say £10 per point.Now consider the following scenarios.

QuestionScenario 1

At the close of the day’s trading, the FTSE stands at 5080.What is the profit/loss if you adopted Choice A?What is the profit/loss if you adopted Choice B?

Outcome A:Having bought at 5030, you correctly predicted that the spread

was too low.For every point over 5030 you win your unit stake, i.e. £10.Therefore, you win (5080 - 5030) £10 = £500.

Outcome B:As a seller at 5020, you incorrectly predicted that the spread was

too high.For every point over 5020 you lose your unit stake, i.e. £10.Therefore, you lose (5080 - 5020) £10 = £600.

39

SPREAD BETTING

Risk ManagementWhether you are selling or buying, you can indicate to traders alevel at which you would like your bet to be closed without furtherreference. This is known as a stop loss, and it is designed to mitigaterisk.

ExampleYou ask, in May, for a price on the June FTSE 100, i.e. the closing

price on a stipulated day in June.You are given a June FTSE 100 quote of 5310 – 5320.Your view is that the market will rise, so you buy at 5320, for £10

a point.As it is an unpredictable market you decide to take out some

‘insurance’. To do this you place a stop loss order with the tradersat, say, 5270.

This means that as soon as the quote on June FTSE 100 falls to, orbelow 5270 – 5280, your bet will be closed at the first tradingopportunity, for a loss of (5320 - 5270) £10 = £500.

This is a protection against further losses if the market continuesto fall. However, the stop loss cannot be guaranteed at the preciselevel requested, particularly in fast-moving markets.

Guaranteed Stop LossThis type of order offers complete protection at the chosen stop

loss level, but involves an extra premium. The service can be appliedwhen buying or selling.

ExampleYou ask, in May, for a price on the June FTSE 100, i.e. the closing

price on a stipulated day in June.The quote is 5350 – 5360.You think that the level is too high, and therefore you sell at 5350,

for £10 per point.However, you decide to leave a guaranteed stop loss order at 5450.

For this there is a premium charged of, say, 3 points. Hence, theselling level is now at 5347 (5350 - 3).

Now, assume that between the date of the agreement and its expiry

40

BETTING TO WIN

(in June) some very optimistic financial news is released overnight,during the close of trading.

The consequence may be that the spread for the June FTSE 100index opens sharply up at, say, 5510 – 5520.

Without GSL (i.e. with a standard stop loss order), your loss wouldhave been:

(5520 - 5350) x £10 = £1700.With GSL, your position is automatically closed at your

guaranteed level of 5450 (not 5520).So, you sold at 5347 for £10, and the bet is closed at 5450.Therefore, your loss = (5450 - 5347) £10 = £1030.This limits the loss on your bet to £1030.

Limit OrdersYou can place a limit order at the same time as you open a bet, or

anytime thereafter, in order to close the bet if it reaches a pre-determined profit level.

ExampleYou buy, in May, the June FTSE 100, at 5270, for £3 per point, in

the belief that the market will rise.However, you think that if the market goes up by 100 points it

might well then fall back, so you place a limit order to sell at 5370 for£3 if the quote reaches 5370 – 5380 to close the bet.

Assume now the spread for the June FTSE reaches 5370 – 5380 afew days later in May.

Your limit order is activated and you sell at 5370 for £3.You win: (5370 - 5270) £3 = £300.N.B. You can also place forward orders to open a bet (buy or sell) if

a market rises or falls to a level you specify.

Managing Risk By Closing The Bet

Closing For A Known Profit

You ask, in May, for a price on the June FTSE 100, i.e. the closingprice on a stipulated day in June.

41

SPREAD BETTING

The spread is quoted at 5520 – 5530. You think that the spread istoo high, and therefore you SELL at 5520 for £5 a point.

Over the next few days, and before expiry of the contract, theFTSE has gradually fallen, so you ask for a new quote. On calling,you are offered a spread of 5460 – 5470.

You buy the June FTSE at 5470 for £5.You have closed out with a sure profit of:

(5520 - 5470) £5 = £250

Closing For A Known LossAssume now that the daily FTSE is standing at 5530 – 5540. Youthink that the spread is too low, and therefore you BUY at 5540 for£15 per point.

Later in the day, you call back and find that the quote has slippedto 5505 – 5515. You are concerned that the market may fall furtherand decide to cut your losses.

Since you originally bought for £15 a point, to fully close theposition you need to sell for £15 a point at the new selling price of5505.

Your sure loss is:(5540 - 5505) £15 = £525.

In both the above cases you have closed the bet by making a betin the opposite direction, of the same stake.

If you make a bet in the opposite direction, for a smaller stake,you have partially closed the bet.

OptionsThis section is intended for advanced study only. Please feel free toskip it if you wish.

Options allow you to choose whether to buy or sell a particularmarket.

There are two types of option – calls (backing the market to rise)and puts (backing the market to fall).

Options offer the security of knowing your exact downside whenbuying a call or a put.

42

BETTING TO WIN

Call OptionsExample

In April, with the FTSE 100 standing at 5000, you believethat the market is set to rise substantially. However, ratherthan bet on the June FTSE futures market, you decide that itis less risky to buy a call option, because your maximumdownside is known in advance.

Say that the June FTSE 5200 call stands at 40 – 50.You decide to buy at 50 (this is the premium) for £10 a point.

Outcome A: Index risesSay, now, that the index rises to 5250 by 1st May. The June

FTSE 5200 call will also have risen, say to 90 – 100.You may now take a profit by selling the option at 90, for a

sure profit of £400, i.e. (90 - 50) £10.Alternatively, you could wait till the expiry date in June

and take your chance that you will be able to exercise theoption to buy at 5200. As long as the market at expiry standsabove 5200 (say 5280) you exercise your option to buy at 5200and win the difference between the closing price (5280) andthe buy price (5200), multiplied by your unit stake. From thismust be deducted the premium (50) times the unit stake.

However badly the FTSE performed, the premium cannotfall below zero and, therefore, the most you could lose bybuying the option at 50 for £10 a point was £500.

Outcome B: Index fallsYou buy a June 5200 call at 50 with a stake of £10 per point.

The index then falls sharply.At the date of expiry in June, the FTSE stands at 5100. You

cannot exercise your option to buy, since your call is worthless(the market failed to rise above 5200), and you lose thepremium (50) multiplied by your unit stake (£10 a point), i.e.£500.

Irrespective of how far the index drops, your loss is cappedat £500 in this example.

43

SPREAD BETTING

Put OptionsExample

In May, with the FTSE 100 standing at 5050, you are offered aquote for a June FTSE 4900 put at 84 – 94. This gives you the right tosell the index at 4900 on or before the expiry date, having paid apremium of 94.

Say, now, that the index falls to 4710. Since your put strike pricewas 4900, you exercise your Option.

The difference between the strike price and the market price =4900 - 4710 = 190.

However, the premium is 94.So your profit = 190 - 94 = 96 x £5 = £480.

Let us now consider the opposite scenario. You buy a 4900 putwith a stake of £5 per point. The index then rises. You do not exerciseyour option to buy, since your put is worthless.

Irrespective of how far it rises, your loss is your stake multipliedby the Option price (in this case 94):

Loss = 94 x £5 = £470.

44

BETTING TO WIN

Chapter 4: Betting On The

‘Exchanges’

If spread betting is the youngster of the betting mediums, so-called‘exchange betting’ is the toddler.

The idea behind betting exchanges is quite simple. It is to matchup someone who wants to bet on something at a given price withsomeone else who is willing to offer that price. This way both sidesare happy, at least until the result is known.

The person who bets on the event happening at a given price isthe backer. The person who offers the price is known as the layer. Ineffect, he is laying the bet in exactly the same way as a bookmakerdoes.

The advantage of this form of betting is that it essentially cuts outthe formal bookmaker, and thereby the bookmaker’s margin. It alsoallows everyone on the exchange to act as a bettor (backer) orbookmaker (layer) at will. Indeed, it is possible to back and lay thesame event.

There are drawbacks, however, to this type of betting, and aprimary one is that the operator of the exchange deducts acommission, normally from winning bets, and usually up to 5%.The other drawback is that there may not be enough peopleinterested in acting as a bookmaker on certain events. Even wherethere are people willing to lay bets, sometimes these are to verysmall sums.

The betting exchanges certainly increase the choice on offer,however, to those interested in making money from betting.

The Way To Bet On The ExchangesThere really is excellent value to be had from betting on theexchanges, but only a very small proportion of those who bet knowhow to use them.

Those who are interested in learning will find it surprisingly simpleonce the initial unfamiliarity has worn off, and great fun.

The idea is that you get to act as a bettor or a bookmaker, or bothat the same time.

45

BETTING ON THE ‘EXCHANGES’

Backing On The Exchanges

ExampleSay you are considering trading the outcome of a match between

Leeds and Liverpool.Your first consideration is whether the odds about a Leeds win, a

Leeds defeat or a draw are in your favour.The major betting exchanges present you with the three best odds

and stakes which other members of the exchange are offering. Forexample, for Leeds to beat Liverpool the best odds on offer mightbe 3 to 1 (4.0 in decimal notation), to a maximum stake of £80, 11 to4 to a further stake of £100 and 5 to 2 to a further stake of £500. Thismeans that you can stake up to a maximum of £80 on Leeds to beatLiverpool at odds of 3 to 1, although you could bet a further £100 at11 to 4 and a further £500 at 5 to 2. These odds, and the amount youcan stake, may have been offered by one or more other clients whobelieve that the true odds are longer than they have offered.

Say you think that the true odds of Leeds winning should be 5 to2, you may decide to stake £80 at the 3 to 1 and a further £100 at 11to 4, since both these odds are longer than you think the true oddsare. If you wish to be cautious, however, you may decide to stake asmaller sum on the outcome.

Before staking the money, however, you should check that nobookmaker is offering as good a deal, or better. This is important,since there is usually a commission deducted from winnings on thebetting exchanges, whereas returns with the bookmakers arededuction-free.

An alternative option available to potential backers is to enter theodds at which you would be willing to place a bet, together withthe stake you are willing to wager at that odds level. This request(say £50 at 4 to 1) will then be shown on the request side of theexchange, and may be accommodated by a layer at any time untilthe event is over.

Laying On The ExchangesTo lay a bet you set the odds you wish to offer and the stake you areprepared to accept at those odds. Say you think it unlikely that Leeds

46

BETTING TO WIN

will win, you may decide to lay Leeds at, say, 3 to 1, instead of backingthem. Perhaps you are willing to lay Leeds at 3 to 1 up to a stake of£20. This means that you are willing to accommodate a backer (orbackers) up to a total of £20 at 3 to 1. If your bet is fully matched, youwill be liable to pay out £60 (3 x £20) if Leeds go on to win, but youkeep the £20 if they lose or draw.

Betting exchanges include Betfair (www.betfair.com), Betdaq(www.betdaq.co.uk, www.betdaq.com), Sporting Options(www.sportingoptions.co.uk), GGBet (www.ggbet.com) and Betsson(www.betsson.com).

Intrade As An ExchangeIntrade is a betting exchange which works in a similar way to futuresmarkets. In particular, standardised contracts are bought and soldbetween exchange members. Members of the exchange trade witheach other, with the operator charging a transaction fee on eachtrade.

All contracts can be bought or sold, and there are two types of contract– short-term contracts and long-term contracts. Short-term contractsusually cover individual events, such as a particular football match.Long-term contracts usually cover seasonal events, such as the totalnumber of points obtained by a team over the course of the wholeseason.

There are two basic methods of trading. The first involves PIX(Percentage IndeX) contracts, the second totals contracts. Let usconsider these in turn.

PIX ContractsPIX contracts are percentage representations of potential outcomes

of sporting events.The outcomes range from a notional 0 to 100. 100 represents a

positive outcome, i.e. a win. Zero (0) represents a negative outcome,i.e. a non-win. The contracts can, in principle, trade anywherebetween 0 and 100. When the outcome is determined, however, thecontract will expire at either 0 or 100.

Multiple PIX contracts may be listed on the same event. Thesimplest case is where there are three possible outcomes , i.e a win,a draw and a loss.

47


ExampleArsenal v. LiverpoolThere are three possible outcomes, and therefore three linked PIX

contracts :

1. Arsenal win2. Liverpool win3. Draw

Thus Arsenal might be quoted at 55 – 58.At this quote, clients can buy the Arsenal PIX at 58, or sell the

Arsenal PIX at 55.Say a client decides to buy. He may than buy, say, 100 contracts at

a price of 58. At this price, he is risking 58 trading units (or ‘ticks’) oneach contract if Arsenal lose or if there is a draw. The reason is thatthe contract makes up at 0 in the event of any outcome other thanan Arsenal win. The potential upside is 42 ticks on each contract ifArsenal win, i.e. the difference between the expiry price (100) andthe buy price (58).

At any point during the course of the market, the client may decideto close the trade by selling a contract previously bought, or viceversa. This might lock in a loss or a profit, depending upon how themarket has moved in-between.

Totals ContractTotals contracts are priced to represent some likely quantity, such asthe total number of goals scored in a football match, or number ofruns in a cricket match. Each totals contract will have a given range,as well as a prescribed tick size and tick value. If an event results inmore or less (e.g. runs, points) than the pre-set range, the contractwill expire at the specified minimum or maximum price.

ExampleMan United Total Season Points:Tick Size 0.1, Tick Value 10p, Min Price Value 50, Max Price Value

100.In this example the range is specified as 50 points to 100 points.

So if a client decides to BUY 10 Man United contracts at a price of

48

BETTING TO WIN

80, and the final result is 82, the trader has a profit of 20 points (10contracts of 2 points) or 200 ticks (tenths of a point). Since each tickin the example has a specified value of 10p, the profit is £20 (200ticks at 10p per tick). The maximum profit is 200 points (10 contractsof 20 points), which would be realised if Man Utd amassed 100 ormore points.

The Tradindex Sports betting exchange also works on a 0/100principle. Bets are placed with other clients of the exchange, in theform of shares which trade between 0 and 100. Winning team sharesexpire at 100. Losing team shares expire at 0. In a draw both teams’shares expire at 50.

Perhaps the most interesting manifestation of this sort of exchangeis that offered by Tradesports (www.tradesports.com) where traderscome together to exchange contracts based on their opinions aboutthe likely outcomes of a diversity of topical issues and news stories.

49


Chapter 5: Bookmakers And The

Exchanges

Comparing The Options

2002 Cheltenham Festival, Tuesday, March 12thThere is so much tax-free competition between bookmakers thatthe modern-day bettor is faced with a paradise undreamed of in thenot too distant past. Now add in the option of comparing the offersfrom the bookies with the betting exchanges, and it is not difficultto fathom why the odds are starting to look loaded in favour of thebettor. The trick lies in using these options to maximum advantage.

Take as an example the first day of the 2002 Cheltenham Festival.Six races featured on the card, including the Supreme Novices’Hurdle, the Arkle Trophy and the Champion Hurdle.

Diehard followers of the Racing Post advice forums will haveturned straight to the Pricewise page in large numbers, for TomSegal’s idea of value. Win or lose, the conventional wisdom statesthat the prices advised in this column will not usually last long, andso if you want to choose the selected horse/s, you will need to do soearly.

On Champion Hurdle day, there were five new pieces of advice.

2.00 Cheltenham: Gerrard Supreme Novices’ HurdleThe first new advice was Mutakarrim in the 2.00, advised at 50 to 1with Victor Chandler, for a stake of 2 points each way. The next bestprices available about the Dermot Weld-trained 5-year-old were the40 to1 available with Blue Square and Surrey Racing, both identifiedin the Pricewise odds comparison table. Visitors to the Oddscheckersites were alerted to the additional 40 to 1 available with Betabet,though not pointed in the direction of Surrey. Elsewhere there weregeneral 33s and 25s on offer.

Clearly, Chandler was the firm to go with. There was a slightproblem, though, in that Victor C’s operation had opened up at theunusually early time of 8 am, just for the duration of the Festival.

50

BETTING TO WIN

Those unaware of the revised opening hours were, of course, unableto grab the price, which lasted as long as one might expect in theface of the astute early birds just waiting for such an unexpectedopportunity. Those seeking solace in the 40 to 1 available with BlueSquare were accommodated for longer, but to a maximum of just£10 each way.

Still, dragging around the various outlets allowed a reasonablesum on the Pricewise fancy at 40 to 1.

Whatever price you took, and especially if you got the early 50 to1, there might be money to be made regardless of the outcome. Howso? Well, this is where the betting exchanges come in.

Take Betfair as an example. By 9.10 am, it was possible to backMutakarrim at 35 to 1 to the extent of £90, at 34 to 1 for a further £2and at 33 to 1 for another £153. More interestingly, it was also possibleto lay the horse at 37 to 1 for £5 or at 39 to 1 for £25. If you were of amind, there were people out there wishing to stake another £100with you if you would offer them 49 to 1. If you had bet £30 on thehorse at 50 to 1, of course, you could now lay the horse at 39 to 1and shorter to the same stake. If the horse won, you would still bein profit. To be precise, your £30 on Mutakarrim at 50 to 1 withChandler would have yielded you a tax-free £1500. In debit, youwould have had to pay out at 39 to 1 those punters who staked £25with you (a total of £975), and at 37 to 1 to those punters who staked£5 with you (a total of £185). This makes a grand total of £1160.

Your net profit on the deal works out at £340 (£1500 - £1160).If the horse lost, you would have lost the £30 you staked with

Chandler but gained the £30 staked with you by other clients of Betfair.A risk-free £340, then? Not quite, because of the 5% commission

charged on winnings to all but the very regular or high-rollingplayers on the Betfair Exchange (who pay a lower commission). Inreality, you stood the chance of winning £340 for the potentialdownside (if the horse lost) of 5% of £30, i.e. £1.50.

In order to realise an actual risk-free profit, you would haveneeded to lay the horse on the exchange for a little more than youhad staked on it to win.

For example, let’s say you staked just £25 on Mutakkarim at 50 to1 with Chandler. In this case, you would stand to win £1250 if thehorse won.

51

BOOKMAKERS AND THE EXCHANGES

However, by laying the horse at 39 to 1 (for £25 – as before) and at37 to 1 (for £5 – as before), you would be obliged to pay out £1160 toyour Betfair colleagues. That makes a net profit of £90. This is ratherless than before, but at least it is risk-free. Why so? Well, if the horselost you would lose £25 to Chandler, but you would gain £30 fromyour Betfair dealings, which is still £28.50 after deducting the 5%commission. The difference between your winnings on Betfair andyour loss to Chandler is £3.50 (£28.50 minus £25).

In other words, if Mutakkarim wins, you win £90. If Mutakkarimloses, you win £3.50.

A free lunch every time? Not really, as the evidence of the secondPricewise horse to be advised proves so tellingly. The horse inquestion was the Noel Meade-trained 6-year-old, Scottish Memories,in the very same race, and also tipped at 50 to 1. Sounds similaradvice, but as a betting proposition, we were talking a whole differentball game. The first warning sign was the fact that Victor Chandlerwere still willing to offer the horse at 50s long after the 8 am openingbell, and it was possible to take the same long price with the likes ofWilliam Hill, Ladbrokes and Surrey, as well as SportingOdds andSports.com.

Clearly, the, the 50 to 1 did not represent such unique value. Still,according to the Pricewise column, it represented value nevertheless.The advice was to take the price for 2 points each way.

Those unsure of the value might have looked on a shrewdly placed£20 each way, say, as a solid bet anyway. The reason for theiroptimism lay in the hope that the odds would plunge from 50 to1simply because it was the Pricewise fancy, and that the opportunitywould soon open up to hedge, or lay, the horse at rather less than50 to 1. Suitably planned, a risk-free bet on Scottish Memories lookeda banker.

The reality was rather different, as would-be hedgers in theexchanges might care to bear in mind. How so? Well, it is true thatfor a brief few minutes there were those on Betfair who were willingto stake £14 at 49 to 1. Those waiting for the gravy train were to besorely disappointed, however, as it never reached the edge of thestation. Instead, by 10.15 all that was available to the layer was alonely offer of £3 if you would give odds of 59 to 1, and a further £9floating around in pursuit of 65 to 1. Worse still, clients of Betdaq or

52

BETTING TO WIN

GGBet were seemingly unwilling to make an offer to back thewretched thing at any price.

Anyone would be forgiven for thinking that the horse had beenspotted limping to the course.

In the event, Scottish Memories performed rather well, finishinga gallant seventh at 50 to 1 despite being severely hampered as hewas making his challenge. Who knows how close he would havegot with a bit more luck? So much for listening to the market.

Mutakarrim, for the record, finished a poor 18th, at a starting priceof 25 to 1.

The other end of the market told another story.The favourite at the opening of market hostilities was Like-a-

Butterfly, available at a best-priced 5 to 2 with Blue Square, Surreyand Sportingbet.

Subsequently, the best odds at which it was possible to place abet on Like-a-Butterfly on the exchanges was 2.6 to 1 (decimal 3.6),which is marginally longer than the bookies’ best price, but subjectto the penalty of a 5% commission on any winning bet.

By the time the market opened on-course, the generally availableprice was 2 to 1, and the starting price was as short as it ever got, at7 to 4. Despite this, one intrepid punter managed to place a bet of£100,000 to win £225,000, and two other bets of £25,000 were placedto win £56,250.

Like-a-Butterfly won by a neck!

2.35 Cheltenham: Irish Independent Arkle Challenge Trophy ChaseThe third Pricewise tip of the day was Youlneverwalkalone in the 2.35.Trained by Christy Roche, the 8-year-old was advised at 14 to 1 withSunderlands, to a proposed stake of 3 points to win and 1 to place,and Sunderlands were happy to lay the price. 12 to 1 was the bestprice available elsewhere.

This was a horse that it was easy to lay at 12.5 to 1 for most of themorning on the exchanges, for a sure profit if appropriate stakeswere employed. Despite a bet of £4000 each way, and another of£3000 each way, at 10 to1, the horse trotted home last of the finishers.

3.15 Smurfit Champion Hurdle Challenge TrophyNo Pricewise tip for the Champion Hurdle, but three (theoretically)

53


fairly equally matched horses made for a fascinating bettingproposition. The three contenders at the early morning openingwere Valiramix, Istabraq and Landing Light.

Valiramix was a best-priced 9 to 4, available in quite a few places;Istabraq opened in the morning at a best 3 to 1, with UKBetting andSurrey; 7 to 2 could be had about Landing Light with UKBettingand Sports.com

By 9.30 am the exchanges were already accommodating backersand layers of all three to fairly sizeable sums, with Valiramix theclear favourite on Betfair (available at 2.2 to 1 for £2798). Istabraqwas available at 3 to 1 for £1306 and Landing Light at 3.4 to 1 for£265.

By 10.30 it was possible to take all three at longer prices. Valiramixwas on offer at 2.4 to 1 for £545, Istabraq at 3.1 to 1 for £84 andLanding Light at 3.5 to 1 for £899.

By noon the market had settled, and a comparison of the pricesavailable to back and lay on two leading exchanges revealed thefollowing:

Betfair Betdaq

Valiramix

To back:

2.4 to 1 (£2759) 2.3 to 1 (£2263)

2.3 to 1 (£2262) 2.2 to 1 (£1182)

2.2 to 1 (£2759) 2.1 to 1 (£476)

To lay:

2.5 to 1 (£1335) 2.4 to 1 (£1003)

2.6 to 1 (£1265) 2.5 to 1 (£3750)

2.8 to 1 (£8) 2.6 to 1 (£1706)

Istabraq

To back:

3 to 1 (£655) 2.9 to 1 (£928)

2.9 to 1 (£958) 2.8 to 1 (£1429)

2.8 to 1 (£777) 2.5 to 1 (£400)

To lay:

3.1 to 1 (£695) 3 to 1 (£300)

3.2 to 1 (£1731) 3.1 to 1 (£4141)

54

BETTING TO WIN

3.3 to 1 (£681) 3.3 to 1 (£165)

Landing Light

To back:

3.5 to 1 (£189) 3.5 to 1 (£190)

3.4 to 1 (£851) 3.4 to 1 (£921)

3.3 to 1 (£1030) 3.3 to 1 (£574)

To lay

3.6 to 1 (£226) 3.6 to 1 (£604)

3.7 to 1 (£1232) 3.7 to 1 (£2960)

3.8 to 1 (£45) 4 to 1 (£2260)

On course, Valiramix opened at 9 to 4, and started at 3 to 1.The starting price was as long as it was ever available on the day,and the casual punter who took the SP got better odds than allthe sophisticates who had eagerly scanned the early prices andexchanges for value.

Not so in the case of Istabraq, who opened on course at 5 to 2and started as short as he ever was, at 2 to 1.

Landing Light opened at 7 to 2 and started at 100 to 30.The most notable feature of the market movements was that

backers of Istabraq would never have done better than the early7 to 2 available with a couple of bookmakers. Supporters ofValiramix would, in contrast, have been best advised to wait aslate as possible before putting down their money, or else justtaking the starting price. Which all goes to show that there is nohard-and-fast system for always obtaining the best price. Thereare ways of improving your chances, though, and that is, in part,what this book is about.

In the event, those backing any or all of the three at the top of themarket went home empty-handed. Valiramix was going easily whenhe stumbled, and tragically had to be destroyed. Istabraq was pulledup before the third hurdle. Landing Light finished fifth.

3.55 William Hill National Hunt Handicap ChaseThe 3.55 was billed as a match between the Philip Hobbs-trainedseven-year-old, Gunther McBride, ridden by Richard Johnson, and

55


the Martin Pipe-trained Carryonharry, partnered by Tony McCoy.At the opening of the market on the morning of the race, the best

price available about the Hobbs horse was 7 to 1, with Sportingodds,while Carryonharry was available in a place (Betabet) at 15 to 2.

Both these prices were better than was available on Betfair at 9.30 am,where the best price available about Gunther McBride was 6.6 to 1 (for£145), and the best about Carryonharry was 7 to 1 (for £256). Both thesewere consistent with the generally available prices at the time, albeit anywinnings on the exchange were subject to a normal 5% deduction.

Those seeking to hedge the horses for a riskless potential profitwere able, by the time the market had settled down at 11.30, to layGunther McBride (with Betfair) at 6.6 to 1 for £22, and at 6.8 to 1 fora further £168. Carryonharry could be laid with Betfair at 6.4 to 1 for£10, for a further £12 at 6.6 to 1, and at 7 to 1 up to £228.

By noon it was still possible to bet £70 on Gunther McBride at 7 to1, and £129 on Carryonharry at 6.6 to 1, with Betdaq.

Both were available at better prices than would ever be offeredon the course.

Gunther McBride opened at 6 to 1, and started as a 4 to 1 favourite.Carryonharry opened at 6 to 1 and went off at 5 to 1. Both lost,finishing 6th and 9th respectively.

4.30 Fulke Walwyn Kim Muir Challenge Cup Handicap ChaseThe fourth Pricewise tip of the day was No Discount. Trained by TedWalsh, the 8-year-old bay gelding was advised at 12 to 1 (availablequite generally), to a proposed stake of 3 points to win.

This was a horse that was never to be longer in the market. Indeed,by 9.30 it was down to 10.5 to 1 with Betfair (for £114), and no betterthan 10 to 1 to anything more than tiny sums for the rest of themorning. There was absolutely no problem in laying the 11 to 1 tomodest sums.

Meanwhile, the market favourite at the opening of business wasThe Bushkeeper, from the Nicky Henderson yard, available at a fairlygeneral 6 to 1. The Bushkeeper lengthened on the Exchanges to abest of 6.6. to 1 at about 10.30, settling down at 6 to 1 thereafter.

On the course, The Bushkeeper opened at 5 to 1, started at 9 to 2,and won comfortably. No Discount showed nothing on the course,and fell two out when going nowhere.

56

BETTING TO WIN

5.05 Pertemps Final (Handicap Hurdle)The final Pricewise tip of the day was Calladine. Trained by ChristyRoche, and with Charlie Swan on board, the 6-year-old bay geldingwas advised at a general 6 to 1 to a 2 point stake.

It was never better on the exchanges. By 9.30 am, the best priceavailable on Betfair was 5.6 to 1 (for £5), at 5.4 to 1 for a further £50,and an additional £391 could be staked at 5.2 to 1. Simultaneously, itwas possible to lay the horse at the original 6 to 1, to stakes up to£248.

By noon, the best price available on both Betfair and Betdaq was5.2 to 1, for £486 and £602 respectively. Layers could take advantageof backers willing to stake £1776 at 5.5 to 1 with Betdaq.

On the course, Calladine was available at a steady 7 to 2 (with 4 to1 in places), and went off at 7 to 2 favourite, after attracting bets of£3000 to win £12,000, £2500 to win £10,000, and four bets of £2000 towin £8000.

Calladine was badly hampered when making his challenge, andfinished 14th of 24 starters.

By the end of the day, none of the Pricewise horses had yielded areturn, though there were some opportunities to hedge at shorterprices than those advised on the exchanges. The on-course andstarting prices about these selections were usually significantlyshorter than the morning opening prices.

To a degree, the lessons of Cheltenham are atypical, since the marketis so strong, and the information is so well known prior to raceday. Onmore usual days, a Pricewise tip will often lead to a marked andimmediate shortening of the horse. In these circumstances, it may seemjust a matter of backing the Pricewise horse to whatever the bookie willallow you and then hedging by laying it at a lower price in theExchanges. The danger, of course, is when everyone else is trying todo the same thing. After all, to lay a horse successfully at a given price,someone must be willing to back it at that price, with you.

An example of a successful Pricewise punt along these lines cameno later than the immediately following Saturday, at Uttoxeter. Theselection this time, in the Marstons Pedigree Midlands GrandNational Chase was The Bunny Boiler, advised at a general 8 to 1 (9 to1 with Sunderlands) to a 1 point stake.

57


The favourite, This is Serious, was available at a best-priced 11 to 2.In the event, the money came for the tipped horse, The Bunny

Boiler.By 10.30 am, the best price available about the beast was 7.4 to 1

(to a stake of £59) with Betfair. 7 to 1 was the best available withBetdaq, to £145. Those seeking to hedge could already make apotential risk-free profit by laying the Noel Meade horse at 7.6 to 1(up to £315) with Betfair, and/or at 7.6 to 1 with Betdaq (up to £760).

By 11.30 more money was coming for the Racing Post selection,which was now best-priced with Betfair at 7.2 to 1 (to a total stake of£56) and at 7 to 1 for a further £549.

By 12.30 the best price, available with Betdaq, was just 6.6 to 1 (for£102).

Hedgers were able to lay The Bunny Boiler at 7 to 1 (for £904)with Betdaq, and it was even possible to lay it at 6.5 to 1 with GGBet,albeit to a single offer of £5.

The flow of money for the Pricewise horse was not matched inthe enthusiasm for any of its rivals, and certainly not in the moneychasing the favourite, This is Serious. Opening at a best-priced 11 to2, it was soon available at 6 to 1 in the exchanges to sizeable sums,and longer still once the market had settled.

On the course, offers about This is Serious were less generous,however, and it maintained a steady 9 to 2, before starting at thesame price. The Bunny Boiler duly opened at 6 to 1, shortened to itsshortest ever 5 to 1 and trotted home, an easy 20-length victor.

The lesson of the week was that followers of Pricewise selectionsare usually best advised to get stuck in early, and to try to hedge inthe exchanges once the market has settled down if so inclined. ThePricewise tip is also likely to be a better bet for those seeking to beatthe starting price, or to hedge in the exchanges, if the market is notoverly strong, and if the price on offer is not too generally available.Do not expect, though, to be allowed to stake as much as you wantin such circumstances, or think that you will always be able to hedgeyour bet later in the day. Racing, like life, just ain’t that easy.

Punters Beware!The idea of betting with the bookmakers and hedging in the bettingexchanges, highlighted above, exposes one potential trap into which

58

BETTING TO WIN

the unwary punter can so easily tread. This is courtesy of the‘palpable error ’ rule of which the slapdash bookmaker is so fond.An example, from my own recent experience, may best illustratethe trap.

On 11 May 2002, at 1.42 pm, I received the following e-mail froma source who shall remain anonymous.

‘UKBetting.com have priced the Leeds v. Middlesbroughmatch incorrectly. For a 0 - 0 score they price it up as 10/1 –however, for no goalscorer they price it as 50/1. Take advantageof this … if you want you can lay off the bet at Betfair (0 - 0 iscurrently available at 17/1 to lay) producing a guaranteedprofit.’

Ever the investigative reporter, I placed £50 on ‘No Goalscorer ’ at50/1 with UKBetting, but refrained from exposing myself to thepotentially heavy liabilities I could have incurred by laying 0 – 0 at17 to 1 in the exchanges.

The reason for my caution lay in the ‘palpable error ’ rule.According to this rule, written into bookmakers’ rules andregulations, prices set in error can be rescinded if the error is‘palpable’.

Turn now to Rules 14 and 15 of the UKBetting code for the detail.

‘Rule 14: It is your responsibility to ensure that details ofyour bets are correct. Once you have confirmed your betand it has been accepted by UKBetting.com then changesor cancellation will not be permitted.Rule 15: UKBetting.com will not allow nor acceptresponsibility for any palpable errors or omissions in respectof the posting of prices, elections (runners, players, teams,etc), times or results despite every effort to ensure completeaccuracy. Notwithstanding Rule 14 above, we reserve theright to correct obvious errors (for example an incorrectprice or voiding bets struck after an event is underway).The term ‘incorrect price’ means an error in inputting orreversal of the odds. In such cases, the client will be notified

59


and offered the option to cancel the bet. Where noinstruction is received relevant bets will be settled at theircorrect odds.’

Now the danger is all too clear. Hedging the mythical 50 to 1 at that17 to 1 in the exchanges is a potential road to the poorhouse.

Not wishing to place £50 on No Goalscorer at 10 to 1 (I could getit at 12 to 1 elsewhere if I was interested), I sent off the following e-mail to Customer Services at UKBetting (I could find no telephonecontact number), timed at 2.17 pm.

‘I have just placed a bet of £50 on No Goalscorer in the Leeds/Middlesbrough game, at 50 to 1. Starts 3 pm today. I wouldnot wish to place this stake if you are to invoke a palpableerror rule on me if the bet is successful. Can you clarify thesituation?’

At 2.38 pm came the response:

‘Dear Sir,All bets on No Goalscorer at the incorrect price of 50/1 havebeen voided (at 14.30) as this was a palpable error. The priceshould have been 10/1 as reflected in our Correct Score market.Your stake will be returned when the First Goalscorer markethas been settled, at the conclusion of the game.’

Those who had laid the 0 – 0 in the exchanges at 17 to 1, in theexpectation of a riskless profit, were left biting their nails and beyond.

One other point. 0 – 0 is not exactly the same as No Goalscorer.The difference is the way that ‘own goals’ are counted. Theycontribute to the scoreline, but not usually to the No Goalscorermarket. 10 to 1, say, about No Goalscorer is to this extent better oddsthan 10 to 1 about a 0 – 0 scoreline, inasmuch as a 0 – 0 scoreline betwould lose in the event of an own goal.

This is only a general guideline, however. You should check theindividual bookmaker’s rules to be sure in any particular case.

Part Two: Beating The Bookmakers

Chapter 1: Theory And Evidence

The key factor to consider when considering a profitable approachto betting with a bookmaker is that the bookie offers prices aboutmany more eventualities than you are willing to bet upon. In otherwords, you have the luxury of picking and choosing between allthe prices on offer. Similarly, you have the choice of picking amonga significant number of bookmakers, all of whom are competing foryour custom. Sometimes these prices are so disparate that it ispossible, indeed, to bet on every possible outcome with differentbookmakers and win whatever the result. Such circumstances donot last long, though, as there are plenty of people willing to swoopon such prices, with the effect of forcing them down to a lessgenerous level.

There are, however, many less fleeting ways of tackling thebookmaker which can, if handled carefully, offer the opportunity ofturning the odds in your favour.

Economists and statisticians have worked for over half a centuryseeking to identify these methods, through an examination of whatare known as betting market ‘inefficiencies’ or ‘anomalies’. What ismeant by an ‘inefficiency’ in this sense is an opportunity to make asuperior return by the use of a defined system or approach. Thereare a number of these inefficiencies, and we shall consider themeach in turn.

First, though, we shall turn to perhaps the most well known, andcertainly the most rigorously tested of them all. This is the so-called‘favourite – longshot’ bias.

The Favourite – Longshot BiasThe idea of a so-called favourite – longshot bias at the racetrack wasfirst identified in 1949 by Richard M. Griffith, a psychologist basedat the Veterans Administration Hospital in Lexington, Kentucky.

Griffith examined the odds available about thousands of horsesrunning at US racetracks, and catalogued where each of thesefinished. The results of his research were astonishing at the time,and were soon accepted for publication in the American Journal ofPsychology. What he found was that the shorter the odds at which a

64

BETTING TO WIN

horse started the race, the better on average the value. In otherwords, those who systematically bet on the favourite (the horse withthe shortest odds) would over the long-term win more, or at leastlose less, than those backing any other horse or horses in the field.

This was a startling discovery, because it suggested that it waspossible to earn above-average returns by following a simple bettingsystem, which required no knowledge of anything other than theavailable odds.

This discovery was significant not only for horse race bettors(obviously) but also for economists. After all, how could this loopholeexist?

To the betting public, and that included quite a few economistsand statisticians, the most important thing to find out was whetherGriffith was right, and this meant collecting lots more data, fromdifferent racetracks and at different time periods. Amazingly, studyafter study came up with the same findings. The favourite – longshotbias is indeed real.

In the subsequent fifty years and more, only one significant USinvestigation has indicated otherwise, and that solely for the case ofone atypical, relatively small US racetrack.

In a way, this is not surprising, since laboratory experiments datingback to a classic study by Preston and Baratta in 1948 all point in thesame direction. These experiments all found evidence of a systematictendency by subjects (under controlled conditions) to relativelyunderbet or undervalue events characterised by high probability(short odds), and to relatively overbet or overvalue those with lowprobability (long odds).

Wayne Snyder, writing in 1978, surveyed all the work to that datewhich had looked at US racetrack betting, and concluded that therewas indeed a strong bias which made bets on favourites much bettervalue than bets on longshots. The problem was that the bias wasnot big enough to cover the deductions from bets levied by theoperators of the American Tote-only (‘parimutuel’) system.

A subsequent classic US study, undertaken by Richard Thaler andWilliam Ziemba in 1988 was a little more optimistic (from the punter ’spoint of view). While they confirmed that track deductions weretoo large to make bets on the aggregate of short-priced favouritesprofitable, they found that such a strategy was profitable at odds of

65

THEORY AND EVIDENCE

3 to 10 or shorter, i.e. very short odds. 3 to 10 means that a punterwould have to stake £10 to win £3, with the stake to a winning betreturned.

The problem with the US studies is that they are confined to thesystem prevalent in that country, i.e. a Tote monopoly.

Those seeking to turn an honest profit from their betting activitiesin the UK faced a very different environment. Both bookmakersand the Tote were available on-course, and competed for the punter’smoney. It was not long before attention was turned to the existenceof a favourite – longshot bias at bookmaker odds.

A pioneering study was undertaken in 1951 by E. Lennox Figgisin his book, Focus on Gambling. This and subsequent studies by Figgis,were to become so influential that they were quoted in the finalreport of the 1978 Royal Commission on Gambling. The favourite –longshot bias was alive and well in the UK, at the odds quoted bybookmakers.

The way that Figgis conducted his studies was to collect data onthe starting prices of the horses. The SPs, as they are called, are theindependently determined assessment (by professional assessorsat the track) of the general price at which a bet could have beenplaced on a horse with course bookmakers at the start of the race.Most off-course bets are settled at this price.

Figgis’ evidence on SPs collected for races run in 1950, 1965 and1973 demonstrated that for the shortest odds examined, i.e. 2 to 5(£5 to win £2, with the stake to a winning bet returned), the averagepre-tax return varied from 97.2% in 1950 to 108.1% in 1965 to 108.5%in 1973. Calculations of the returns in 1975 and 1976, performed forthe Royal Commission on Gambling, found rates of return of 112.1%and 107% respectively. In other words, in four of the five yearsexamined, betting on all horses starting at the shortest odds examinedwould have yielded a very healthy pre-tax profit.

To pay for this healthy profit, it is inevitable that other puntersare losing in an unhealthy way. If they were not, the bookies wouldbe out of business. The big-time losers, Figgis found, were thosewho backed the longshots (horses starting at long odds). Hiscalculations for the longest odds range, i.e. 20 to 1 and longer, provedthe point. The average returns varied from as little as 23.8% in 1950to 37.3% in 1965 to 23.2% in 1973. These low returns to longshots

66

BETTING TO WIN

were in line with the USA findings, although much morepronounced in extent.

Over all odds ranges, the average return was about 80%, i.e. a20% loss to all money staked.

It was, however, Jack Dowie who first brought these ideas to thenotice of the academic world with a path-breaking article publishedin the journal Economica in 1976.

Dowie calculated the expected return to bets paced on horsesstarting at each of a wide variety of starting prices for the 1973 flatseason. His sample of 2777 races revealed evidence of a significantlongshot bias, to the extent that a pre-tax profit could have beenmade by betting on all horses starting at 4 to 6 or shorter. Again, thereturn to longshots (especially extreme longshots) was far worse inextent than that reported for US racetrack betting markets. The samestudy also showed that bets, to level stakes, on all horses returnedat 1 to 2 or shorter would have provided a pre-tax profit of 9.3 percent. That’s to a blind strategy of betting on each and every case,and doesn’t even allow for the fact that it is often possible to beatthe starting price.

Robert Henery was another to wade in, with his examination,published in 1985 in the Journal of the Royal Statistical Society, of 883races run in 1979 and 1980. The average return to a unit stake wascalculated over various odds ranges, demonstrating a return of 97.9%to bets on all horses starting at 2 to 5 or shorter, but only 10% tohorses starting longer than 38 to 1.

Confirmation of the findings of the academic studies wasprovided, interestingly enough, in the Ladbrokes Pocket Companion,Flat Edition, for 1990. The findings covered the flat racing seasonsfrom 1985 to 1989. As expected, bets at shorter odds provided muchbetter returns than at longer odds. Indeed, the results suggested apositive rate of return to a strategy involving the consistent placingof bets on horses which started at odds of 1 to 2 or shorter. Of 344horses starting at odds between 1 to 2 and 1 to 5, 249 won, giving aprofit of 0.52% if the same stake (level stakes) was placed on everysingle runner in this category. There were 96 examples of horsesstarting as even hotter favourites (between 1 to 5 and 1 to 25). Ofthese, 88 won, for a level stakes profit of 6.5%. Of the 35 favouriteswho went off at odds of 1 to 8 or shorter, all won.

67

THEORY AND EVIDENCE

At the other end of the spectrum, longshots starting at 25 to 1won just once out of 52 occasions, and level stake bets on all horsesstarting between 25 to 1 and 100 to 1 would have produced a loss of67.76% of everything staked (43,426 runs, just 441 winners).

The following table, reproduced from the Ladbrokes PocketCompanion, clearly shows the nature of the bias against longshots.

Odds Wins Runs Lev. stakes % profit

1/5 to 1/2 249 344 + £1 - 80 + 0.52%

4/7 to 5/4 881 1780 - £82 - 60 - 4.64%

6/4 to 3/1 2187 7774 - £629 - 00 - 8.09%

7/2 to 6/1 3464 21681 - £2237 - 00 - 10.32%

8/1 to 20/1 2566 53741 - £19823 - 00 - 36.89%

25/1 to 100/1 441 43426 - £29424 - 00 - 67.76%

Source: Ladbrokes Pocket Companion, 1990 Flat Edition

Before you start trotting around the globe armed with thisknowledge of how to blunt the odds against you, beware of oneplace where all is not what one might expect. I refer to the HappyValley and Sha Tin racetracks of Hong Kong where betting on thehorses seems to be not so much a hobby as a mad passion.

Studies, covering 5343 races, published by Kelly Busche andChristopher Hall (1988), have shown that the traditional favourite –longshot bias just doesn’t exist over there. The same also seems togo for Japan. Why so? Do all the shrewdies live in the Far East, ormake their way there on the first available flight? Nobody knowsfor sure, although theories abound. One of the most popularexplanations is that the Tote pools are so big that it pays professionalgamblers to set up shop with the most sophisticated data processingmodels. They then use these models to mop up the money placedin the pool by mug punters so silly as to overbet the longshots. Thisbrings the odds into line with the true probabilities, and eliminatesthe bias. Unlike bookmakers, the pool operator is happy to pay themand pay them again, using the money put into the pool by less astuterace fans.

So much for Hong Kong. Over here, things are very different,and no doubt the majority of punters will continue for some time to

68

BETTING TO WIN

line the bookies’ pockets, taking that 20 to 1 about the 50 to 1 chance!So much for the horses, but how about the dogs? To answer this,

you need look no further than the work of Michael Cain, DavidLaw and David Peel, who demonstrated the same bias at the dogtrack in a 1996 study. Because of the size of the bookmakers’ margins,however, they were unable to translate their findings into a strategycapable of yielding positive profits.

So the evidence is overwhelming. A consistent strategy of bettingon horses starting at shorter odds will yield a better return thanbetting at longer odds. Indeed, for the UK at least, and in at leastone study in the USA, it appears that a policy of betting all theshortest odds good things could, over certain past seasons, haveyielded a pre-tax profit.

Academics differ about what causes the affliction to longshot-backers, but all are agreed that it exists. What isn’t so welldocumented is whether the same applies to football.

Only two published studies looked specifically and uniquely atfootball in this context, the most recent being published in theFebruary, 2000 edition of the Scottish Journal of Political Economy.Despite the title of the journal, the investigation is based on theEnglish Premier League. This study, entitled ‘The favourite –longshot bias and market efficiency in UK football betting’ providesconvincing evidence that what applies to racing applies also tofootball.

The key finding is that the odds available with fixed-oddsbookmakers about very short odds-on favourites provides asignificantly superior return on average than do the odds availableabout longshots. In other words, the traditional favourite – longshotbias is alive and kicking on the football pitch, which means that ifyou know nothing else about football, your best bet is to back thefavourite.

Among correct score odds, the best value, according to the study,appears to lie in backing very short-priced favourites to win by ascore of 1 – 0, 2 – 0, 2 – 1 or 3 – 2.

These findings confirm an analysis written in 1998, co-authoredby myself and David Paton. That analysis, published in the journal,Applied Economics Letters, suggests that the longer the odds about ateam playing in an English Premiership football match, the lower

69

THEORY AND EVIDENCE

the return, on average, to a bet on that team, and that this patternpersists at all odds level. In other words, when betting on football,shorter odds tend to represent better value.

That sums up the evidence for horse racing, greyhound racingand football, but what about all the other sports out there? Well,fortunately for us, Cain, Law and Peel have yet again come to therescue with an article in the Bulletin of Economic Research.

Their data set consisted of the odds and results for 50 boxingmatches, 132 Sunday League cricket matches, 647 snooker matchesand 91 tennis matches at the Wimbledon championships. For goodmeasure, they also looked at 24,603 US baseball games.

Their results confirmed the existence of the usual bias in favourof backing favourites for all the other sports, including baseball.Indeed, a system based on backing strong favourites (roughlydefined as 1 to 2 or shorter) would have generated, in the case ofboxing and cricket, pre-tax profits ranging from 12% to 16%.

There you have it then – the evidence from a series of studiesspanning half a century. If you know nothing else, a blind strategyof betting on favourites would have yielded a much better returnthan betting on any other outcome, and the shorter-priced thefavourite the better. Indeed, in some studies, such a strategy, at veryshort odds, would have produced a significant profit.

Be careful, though, before sprinting off, armed with all thisinformation, to back a series of short-priced favourites. Statistics,after all, work only in the long-run, or on the average. In the longrun, of course, we are all dead. The art of turning the advantageyou may have gleaned from the evidence you are starting toaccumulate will be the subject of our next chapter.

70

BETTING TO WIN

Chapter 2: How To Bet When The

Odds Are In Your Favour

The previous chapter showed how, by betting on short-oddsfavourites, you can in certain circumstances turn the odds in yourfavour. Much of the rest of this book will be about helping you toidentify other ways of turning the odds in your favour, or identifyingthose times when they already are. This is the concept of value, andit has never been more prevalent than in the cut-throat world ofmodern bookmaking.

In this chapter, we will look at the best approach when you spotvalue. In particular, how do you turn the advantage into securelong-term profits? Note the two words – secure and profits – in otherwords, risk and reward.

An example will explain.Let’s say that a kindly bookmaker offers you 11 to 10 about a coin

landing heads up. This offer means that if you call heads you win£11 if the coin turns up heads, but you lose £10 if it turns up tails.Assuming it’s a fair coin, you might not be too averse to a piece ofthe action. After all, half the time you can expect to win £11 and halfthe time to lose £10. Over a period of time, you cannot but make aprofit.

The problem of course, is that what will happen in the long rundoes not always happen right away. Indeed, it might be a long timeindeed before you are able to turn the theoretical edge in your favourinto a deeper bulge in your money pocket than you started out with.

That, of course, is the way with the real world. You just can’ttrust it to behave as it should, however much the odds are shadedto your advantage. As I said before, in the long run we’re all dead,and unless we’re careful we may go bankrupt on the way!

Of course, the generous bookmaker will, on average, lose at hisquote of 11 to 10, but along the way your bankruptcy could spell hisfortune.

The likelihood of bankruptcy engulfing you before you can turntheoretical into practical advantage depends, naturally, on yourstake. Thus if you bet just a single penny each time an honest coin

71

HOW TO BET WHEN THE ODDS ARE IN YOUR FAVOUR

twists in the air at 11 to 10, you are free from the worry of imminentdoom, but you are hardly likely to get rich in the process.

If, on the other hand, you wager half your gross income becausethe odds are in your favour, you are likely to win rather more, butyou’re also more likely to have that run-in with those unforgivingbailiffs. Just try explaining to them how that 11 to 10 really was verygood value about the flip of a very small coin.

So what should you do when you spot value? Fortunately, JohnKelly has provided us with just the answer. His so-called Kellycriterion tells you exactly how much to stake on favourable bets inorder to maximise your fortune without sacrificing your peace ofmind.

The mathematical details are a little complex, but the essentialconclusion of his analysis is clear enough. The Kelly strategy formaximising long-term average capital growth is to wager aproportion of your assets equivalent to your advantage at theavailable odds.

For example, if you are offered evens about a bent coin comingup ‘heads’, and you happen to know that it has an exactly 60%chance of landing just that way up, what is your advantage? Well,it’s 20%, i.e. a 60% chance of heads (you win) minus a 40% chanceof tails (you lose). The Kelly formula in these circumstances advisesa stake of 20% of your ‘fortune’ on heads. By the way, for practicalpurposes it may be more prudent to substitute ‘betting bank’ for‘fortune.’ The same, of course, applies to any event where youcalculate the advantage to be 20% in your favour.

Now, what would Kelly say about the outcome of a toss of thattwo-headed coin you’ve kept for a rainy day? Well, with a 100 percent chance of a head, and absolutely no chance of a tail, youradvantage in betting heads, however short the odds, is a whole 100per cent. It’s a clear case of betting your ‘bottom dollar ’; it’s mortgagethe house time. After all, you can’t lose, can you? Can you? Yes, youcan! In the real world, even a double-headed coin can turn up tails.Stake accordingly!

72

BETTING TO WIN

Chapter 3: Should You Follow The

Market?

It was available at 20 to 1 with the bookmakers on the morning ofthe race. On the course the first price the bookies offered about itwas 10 to 1. Now, a couple of minutes before the ‘off ’, it stands at 5to 1. This is the classic case of the so-called ‘steamer’.

So you failed to get on at the best price, by quite a margin. Butwhat do you do now? Do you curse your bad fortune (or badjudgement) in missing the ‘value’, or do you knock over the bookies’boards in a headlong rush to grab a piece of the action before thebeast turns an unhealthy shade of odds-on? If only there was a wayof gauging just how much heat is left in these steamers, wouldn’tlife be so much easier?

Enter the professional economist, a big database and a bag ofstatistical tools. The task is then relatively straightforward. First, inputall the early prices on the morning of the relevant races, then all theopening prices in the on-course market, and finally all the changesin the prices before the eventual declaration of an official startingprice. All you then need do is record which horse won.

For the UK, the pioneer of this sort of work was Nicholas Crafts,who in 1985 published an influential article in the journal Economica.The article, entitled ‘Some evidence of insider knowledge in horserace betting in Britain’, generated some fascinating results.

Crafts’ starting point was the idea that there is an element of thebetting population who know more about a particular race than thebookmakers. These people he terms ‘insiders’. These insiders, heargued, were able to take advantage of movements in the oddsduring the course of the day to bet at odds greater than the startingprice. Although this is not an option in a Tote-only system, it is in abookmaking system, where bettors can ‘take a price’ at any timewhile the market is open.

Crafts reasoned that a marked shortening of the odds availableabout a horse during the course of the day may indicate evidence ofinsiders who knew that the probability of that horse winning wasgreater than that implied in the early odds.

73

SHOULD YOU FOLLOW THE MARKET?

He first looked at the prices at which the horses were forecast tostart their race. These so-called ‘forecast prices’ are published in thenewspapers, and Crafts used the forecasts of the leading racingnewspaper of the day, the Sporting Life. He also looked for significantmovements in the on-course market as well.

His data set consisted of 16,769 runners in total, over a periodfrom 11 September 1982 to 8 January 1983. He tested variousdefinitions of a marked movement in the odds, for example adefinition based on the starting price being half as big as the openingprice (5 to 1 from 10 to 1, say). In order to isolate races which mightbe the target of above-average levels of insider activity, Crafts dividedthe sample into handicaps and non-handicaps, the idea being thatthese are distinguished by the amount of established public formavailable about them.

In handicap races, horses are weighted on the basis of past form,which must be established over a series of races, in an attempt toequalise, as far as possible, the chances of all the horses in the race.Such races are less likely, therefore, to offer as much scope for thosein the know to trade profitably as non-handicap races, where theform need not be so exposed to public scrutiny.

Crafts’ results suggest that horses that display a marked shorteningin their odds are indeed characterised by an exceptionally highexpected return at forecast prices. Moreover, this is particularlystrong for non-handicap races.

An examination of the scope for making similar profits duringthe formation of starting price odds in the actual on-course marketproduced similar findings, although splitting the sample as before (intohandicap and non-handicaps) failed to reproduce the earlier result.

Crafts offered supporting evidence in descriptions of bettingpatterns, published in the Sporting Life, about patterns of on-coursetrading. In particular, he identified examples of large sums of moneyplaced on horses with poor previous form, the odds about whichwent on to shorten up considerably and to win.

These results were so significant, if true, that I decided to see ifthey applied to a totally new set of races, for which I personallycollected data. The results of my research were subsequentlypublished in the book, The Current State of Economic Science in 1999.My data set was drawn from all standard flat races in the UK from

74

BETTING TO WIN

March 19 to May 16, 1992, with one additional category added, ofhigher-grade handicaps, which excludes both non-handicaps andlower-grade handicaps. As higher-grade handicaps are the subjectof particular public attention and scrutiny, it would seem unlikelythat they, in particular, offer much scope for the profitable use ofprivate information.

My results were broadly in line with Crafts’ own findings. For thesample of horses that showed the most marked shortening betweenforecast and starting prices (the price halved or more), the expectedreturn at forecast prices was particularly high for non-handicap racesand lower-grade handicaps. The same pattern was found for thesample of races showing a marked shortening between the longestprice available in the on-course market and the starting price.

So one thing seems true for sure. Horses which shortensignificantly in the market are a great bet, at the original price. Indeed,if you don’t make a healthy profit at these prices, you’re doingsomething mighty wrong.

So horses which shorten significantly in the market are a greatbet, at the original price, particularly in certain types of race. If so,what’s the problem? Indeed, is there a problem? Of course there is.The original price has probably long gone by the time you spot whatgood value it was. If only you could ask your friendly layer for the‘best price’, rather than the current (board) price, or the startingprice. But don’t lose heart just yet, for even at starting price thepublished evidence suggests that steamers still represent relativelygood value, and particularly so if you are selective in your choice ofbets. Let me explain.

A horse which shortens is likely to represent a better bet, at anygiven price, than one which does not. For example, a horse whichshortens from 10 to 1 to 5 to 1 represents better value, even at the 5to 1, than an equivalent animal which had always been available at5 to 1. It represents even better value than another horse which hadpreviously been offered at, say, 3 to 1 before drifting out to 5 to 1.That’s the good news.

The bad news is that the final price is, on average, not good enoughto translate this observation into a profitable long-term bettingstrategy. By backing the steamers at their starting price, all you canexpect, on average, is to lose rather less than you would by backing

75

SHOULD YOU FOLLOW THE MARKET?

horses whose odds do not shorten in the same way.You can help your cause, however, if you are selective. To understand

this, we need to turn again to Professor Crafts, and a follow-up articlewhich he published in 1994 in a book called, The Efficiency of RacetrackBetting Markets. In that article, Crafts built upon his earlier work byidentifying a category of horses which had not run for a long time (theseason before last, to be precise). As in his 1985 paper, he used theexistence of a ‘marked’ shortening of the odds about a horse (definedas the forecast price being at least one and a half times as big as thestarting price, say 9 to 1 compared to 6 to 1) to indicate insider activity.He also eliminated from his sample all horses starting at odds of 7 to 1or greater, presumably since these longer-priced horses are on averageless profitable to follow than those starting at a shorter price (see chapter4 for more on this favourite-longshot bias).

The reasoning behind looking at horses returning after a longlayoff is that these are the very horses about which inside knowledgeis likely to be the most valuable. In particular, they may not havebeen seen on a racecourse for months, but someone, somewherehas seen them, and if sizeable sums are going down, you might bewise to pay attention. Moreover, it may be especially significantbecause those who are not ‘in the know’ are unlikely to be riskinghefty sums about these dark entities. So what’s the evidence?

Crafts identified 88 horses over a period between September 1982and November 1987, which had shortened markedly according tohis definition. Backing all of these at forecast prices would have ayielded a rate of return on stakes of a massive 261.9%, and a rate ofreturn even at starting price of a very healthy 55.8%.

The implication of these findings is not difficult to spot. A goodstrategy for those not privy to any inside information about horsesreturning after a long layoff would be to follow the money. Indeed,such a strategy would have doubled your money even at starting price.

The theory makes sense, and the results support the theory. Still,88 observations do not constitute a large sample, at least not in astatistical sense. So before you mortgage the house on the nag thathasn’t been out of the stables since you last cleaned out your localbookie, a word of caution. One study doesn’t make a system. Notnecessarily, anyway. Ultimately, though, it’s for you to decidewhether and when it’s all just hot air.

76

BETTING TO WIN

Chapter 4: Was The Steamer Value At

The Off?

It was the day before Halloween and it provided a nightmare forbookmakers, courtesy of a horse called Tayseer. Normally, asponsored handicap like the Ladbroke Autumn Stakes ShowcaseHandicap at Newmarket is a bookies’ benefit. With 19 runners goingto post, the bookmakers’ notional margin (or over-round, as it isknown) stood at a full 133%. This is equivalent to a two-horse fieldin which one horse is offered at evens and the other at a shadeworse than 1 to 4. Hardly attractive pricing when stated like this,but much less off-putting when spread among 19 contenders. As ahandicap, the form of the horses is relatively exposed, and they aresupposedly weighted so as to equalise as far as possible their chancesof winning. This should make the selection of a winner, in theory atleast, that bit more difficult as well.

Compare the preceding race on the card, the Ben Marshall Stakes.This Class A Listed race featured only eight runners and an ‘over-round’ of just 8%. As a non-handicap there was also no artificialattempt to produce some sort of dead heat.

In the event, the Autumn Stakes Handicap delivered handsomelyto those smart enough to follow the money, and particularly so tothose who placed the early money that was followed. The facts arethese. Tayseer opened in the offices at 5 to 1 generally, and at 11 to 2with one credit bookmaker. On the back of some well-publicisedtips, these prices were quickly taken, and within a very few minutesthe price available on the telephone had plummeted to as low as 2to 1. To limited stakes, the higher prices were still available for awhile longer in the licensed outlets of various bookmakers. The keyquestion remains. Is there still value in the price available at the off?

Tayseer opened on course at 2 to 1, touched 9 to 4 in places, andwas returned at 15 to 8. In retrospect this was a good price, but anyprice is a good price when the horse is past the post and theconnections are celebrating. Whether it was value at the time is aquestion that has been addressed more generally in the precedingchapter of this book. There we came across evidence that these so-

77

WAS THE STEAMER VALUE AT THE OFF?

called ‘steamers’ are particularly good value when they occur innon-handicaps, and/or where the form is not well established orout of date. The value also tends to be concentrated at prices below8 to 1 or so. Of these criteria, the Tayseer phenomenon passed onlythe price test. All we could expect in such circumstances, judged onthe published evidence, is a lower loss on average than if you wereto back horses that have not shortened. Of course, at the originalprice, or even at a range of prices on the way down, the potentialfor a genuine value bet remains. At any price, the key to obtainingvalue is a well-placed belief that the price is set to fall or to fall furtherthan it has already done. On that basis, Tayseer was value, but not,without the benefit of hindsight, at the off.

Fast forward now to the 2002 Guineas meeting, and ‘2000 Guineas‘day.

While most eyes were on the big race, an unusual confluence oftipping interest focused instead on the outcome of the 3.15 atNewmarket, the Class C Labrokes.com Handicap. The interestcentred on the Racing Post’s Pricewise and Betting Bureau columns,both of which highlighted the Akehurst-trained 8-year-old Marsad,carrying second top weight, a tip also featured on one of the biggestprivate forecasting services in the land.

In a 30-horse field, Marsad was on offer ostensibly at 22 to 1 withStanley, and 20 to 1 with Coral and William Hill. Within minutes,the best available was 14 to 1, and by the opening of the market 10to 1 was the going price. Despite a single biggest recorded bet of£300 at that price, the odds duly plummeted to 11 to 2 by the off,thanks in no small part to bookmakers backing the horse on-courseto limit their liabilities.

Marsad flew home and the lesson of this tale is clear. When ahorse is heavily tipped, at least by respected tipsters, it is likely toshorten. If you have missed the price, there may still be value at allprices down to that available at the off. As it happens, this time thenominated horse landed the gamble.

Move forward now just three days to the opening of the Chestermeeting. Similar Pricewise interest led to copycat plunges on twohorses (Northern Desert and The Glen in the 2.25 and 3.25respectively), if not of quite such spectacular proportions. This timethe gambles were foiled.

78

BETTING TO WIN

The reality was the same in both cases. By the off, the price willhave reflected more or less the genuine chance of the beast. At asuccession of prices on the way down the price was probably, onthe basis of known information, on the generous side.

Win or lose, that’s the lesson to be learned for those seeking long-term profitability.

79

WAS THE STEAMER VALUE AT THE OFF?

Chapter 5: The Right Shape

Sometimes an event has a shape about it that just cries out, ‘Bet onme!’ Let me give you an example which provides lessons that canbe used time and again.

The race in question, held at Aintree, at 4.20 on the afternoon ofThursday, 5 April 2001, was the Class A Glenlivet AnniversaryNovices Hurdle for 4-year-olds, a 14-runner race in which two ofthe contenders, Azertyuiop and Bilboa, stood out like shiningbeacons in the pack. At best early prices, you could have picked up2 to 1 about either of them, with a generally available 10 to 1 abouttheir nearest rival. A level pound on each would, as such, haveyielded a net profit of one pound if either won for the risk of a twopound stake, i.e. odds of 1 to 2 about one or other of the two beaconsflashing first past the post.

Conventional betting theory points in one direction in thesecircumstances, and that is in the direction of taking one or other, orboth. The reason lies in the ‘favourite – longshot’ bias, whichindicates that, on average, a bet at longer odds is worse value thana bet at shorter odds. This kicks in especially forcefully at odds above7 to 1 or so, and in larger fields.

Take all this together, and those without special access toinformation, or special skills at processing that information, arelooking for a race with one or two solid favourites, and a host oflongshots. Where better to look than a race like the Glenlivet Hurdle?

I say one or two solid favourites, but ideally we are looking fortwo. The reason for this lies in the weight of evidence which showsthat horses that contract in the betting are particularly good value,even at their new shorter price, while those that lengthen arerelatively poor value, even at their new longer price.

The stand-off between Adrian Maguire, on board Azertyuiop, andThierry Doumen, astride Bilboa, fits the criteria, therefore, forestablishing value. It had what I like to call the right odds shape.

Once the two horses had been identified, to fit the final piece ofthe jigsaw meant a judicious wait for the on-course money to sendits own message. In the event, both horses opened at 2 to 1, followedswiftly by a single bet of £5000 to win £10,000, two bets of £4000 to

80

BETTING TO WIN

win £8000, and three of £2000 to win £4000. All of these were onBilboa. Not a drop of big money was seen on Azertyuiop until hehad lengthened to 5 to 2. Meanwhile, even at 7 to 4 the new favouritecontinued to attract two bets of £2000 apiece.

At the time, of course, those sitting at home had no way ofknowing the exact nature of the trades, but the market told its owntale. By the time that Bilboa had contracted to 7 to 4, the signal to acthad arrived. It would never represent better value, given the availableinformation. Perhaps those cross-checking with the spreads couldhave done even better. Either way, the nod was given Bilboa’s way.

In the event, the French-trained filly never looked in doubt, asshe jumped accurately and effortlessly, before leaving her rivalstrailing in her wake for a 21-length victory.

So there we have it. Two standout favourites in a decent-sizedrace, one of whom eventually attracts the smart money.

In the world of markets and odds, life doesn’t come much moreshapely than that!

81

THE GAMBLER’S FALLACY

Chapter 6: The Gambler’s Fallacy

You’ve tossed a coin and it’s come up heads. You’ve spun the roulettewheel and it’s landed on black. What will happen when you next tossthat coin or spin that wheel? It’s 50 – 50, heads or tails, just as it remains50 – 50 red or black. If the bookie offers you evens, that’s a fair bet.Anything more and you can’t lose in the long run, unless you havegood reason to believe that the coin is bent or the wheel is fixed.

That’s the theory and that’s the logic. All the evidence suggests,however, that real people, and real gamblers in particular, throw allthe laws of probability out of the window when it comes to crunchtime. They suffer from what psychologists and economists like tocall the gambler’s fallacy.

Those who have been exposed to the gambler’s fallacy think thatbecause an event has just occurred, it is less likely than before tohappen again. Take that coin that just landed head side up. Theafflicted now believe that tails is just that little bit more likely nexttime than heads. If they’ve been exposed to a sufficiently strongdose of the fallacy, they might even snub the 11 to 10 being offeredabout heads next flip. The longer the run goes on, the more surethey are that it’s time things changed. ‘Four reds in a row, it’s timefor a black!’ is an all too familiar cry.

The good news for the virus-free is that other people’smisperceptions offer a clear-cut opportunity for gain.

One of the best examples is catalogued by Dek Terrell, of KansasState University, in an article published in 1994 in the Journal of Riskand Uncertainty. Terrell defines the gambler ’s fallacy thus:

‘The “gambler ’s fallacy” is the belief that the probability of anevent is decreased when the event has occurred recently, eventhough the probability of the event is objectively known to beindependent across trials.’

To test for the existence of such a fallacy, Terrell looked to evidencefrom 1785 daily drawings of the New Jersey State Lottery between1988 and 1993. His idea was to study what happened to numbersthat came up twice in a short period of time, specifically a cut-off

82

BETTING TO WIN

period of 60 days. Now, the New Jersey Lottery pays out to winnersa share of the pool of all bets, so a higher payout implies that fewerpeople chose that number. If the gambler’s fallacy applied to the NewJersey Lottery, we would expect the payout to increase if a numberappears twice in a short period of time. This is because people wouldbe less likely, on average, to choose a particular number after it hasjust been drawn. This means a bigger payout if these numbers docrop up because there are fewer winners to share the pool.

Terrell’s results confirmed what advocates of the gambler ’s fallacyhad always thought. Numbers which repeated within a 60-dayperiod had a higher payout than when they won previously on 80of the 97 occasions when this happened. Moreover, the more quicklythe numbers repeated, the greater was the effect. To be precise, Terrellfound that the expected payout on a number that repeated increasedby 28% one day after winning. Therafter it decreased from this levelby about half a per cent each day after the number won, returningto the original payout level (as if nothing had happened) after 60days or so. In other words, the effects of the gambler’s fallacy viruslasted a couple of months in New Jersey.

Interestingly, Charles Clotfelter and Philip Cook, writing in thejournal Management Science in 1993, identified a similar, albeit strongereffect, in the fixed-odds Maryland Lottery.

The fallacy really does seem to get everywhere, at least in thelottery world. How about horse race betting?

Well, just for a moment let your imagination lead you to the track.It is a balmy summer’s afternoon, and you are prowling the paddockin search of the winner of the fourth on the card? The fourth? Yes,you already have your head down. After all, you just knew thefavourite was going to win the first race, but at the last moment thathealthy 8 to 1 about your second choice got the better of you.

Still, how on earth could you have been so stupid as to miss thatgood thing in the second as well? Once – well, accidents happen.But twice – now that’s carelessness. Never mind, you tell yourself,you wouldn’t have won much anyway. Still, that doesn’t explainwhy you missed the too-good-to-be-true 9 to 2 on offer about thefavourite in the third. What now? Well, you really do fancy thatshort-priced animal in the next, and have come ready to turn overthe men with the satchels with one big thump. Hold on a minute,

83

THE GAMBLER’S FALLACY

though, what’s the chance of four jollies in a row? That would beenough to cause a grown bookie to cry – and how often have youseen such a sight? All of a sudden, that longshot doesn’t look suchbad value after all. If you are starting to think this way, you will beable to spot the onset of the first symptoms of gambler’s fallacy. Ifit’s happened before, it’s just that little bit less likely to happen again,or so you think – that’s the nature of the fallacy. Which brings us tothe real question. What is the truth of the matter with respect to arun of winning favourites? Fortunately, Mary Ann Metzger, of theUniversity of Maryland, has done the work, examining bettingpatterns in 12,316 races. Her findings make interesting reading.

Metzger’s study, published in 1985 in the journal, PsychologicalReports, is still the largest investigation to date of this issue, albeitconfined as it is to the US racetrack. In total, she examined 12,316races, her results confirming that racetrack bettors are indeed misledby the outcome of previous races. The single most interestingconclusion of the study is that in those cases where two or threesuccessive favourites won back to back, the betting public tended toavoid the favourite in the subsequent race like the plague. As a result,the odds tended to lengthen about these ‘good things’, making themexceptional value. The simple fact was that punters on the wholewere disinclined to believe that a run of favourites would persist,and so staked relatively little money on that eventuality. This causedthe favourite’s odds to lengthen to a generous degree after a run ofwinning favourites. Similarly, she found evidence that favouritesturned out to represent exceptionally poor value after a run of losingfavourites. In this case, punters were on the whole far too ready tobelieve that it was time for a favourite to turn up, and so put toomuch on the favourite, causing its odds to plummet.

If true, one could do rather worse than wait for a couple of winningfavourites, before neatly stepping in to pick up the 3 to 1 on offerabout that genuine 2 to1 shot.

The truth, it seems, is that favourites really can keep on winning,and the longer it goes on, the better value they become.

A complementary analysis of betting trends looks at a run of winsby dogs emerging from a particular trap. You know the scenario.Trap 4 wins the first race, and the second and the third. Quick, geton the 4 dog, we all cry – or do we? In fact, we do not, just as

84

BETTING TO WIN

advocates of the gambler ’s fallacy would expect. Instead, as shownin a classic study, most of us do just the opposite.

The study in question, published in 1996 in the Economic Journal,by Dek Terrell and Amy Farmer (of the University of Tennessee),focused on greyhounds running at the Woodlands Greyhound Park,in Kansas City. The hypothesis Terrell and Farmer put forward wasbased around a strategy of betting on the greyhound in the trap thatwas occupied by the previous winner. Their reasoning was simple. Ifdog track punters were as open to the gambler’s fallacy as the rest ofthe betting public, one might expect them, on average, to shun thetrap that had just produced the winner. After all, it’s all about under-estimating the likelihood of the same thing happening twice. The testthey conducted was to calculate the return to a strategy of betting onevery greyhound emerging from the trap occupied by the winner ofthe previous race. In the USA the racetrack is Tote-only, so any dip inthe popularity of a particular dog would be sure to translate into animproved payout for those punters going against the crowd.

In a study of 3,795 races at Woodlands, Terrell and Farmer foundthat such a policy was in fact the best strategy available. More thanthat, it was a profitable strategy, yielding a profit of 9% to a policy ofblind obedience to such a simple system, even after deductions.

Will it work for you? Short of a trip across a wide ocean, we can’tbe sure, but it certainly gives us pause for thought.

In summary, then, there is some convincing evidence of theexistence of a gambler’s fallacy, at least in the USA, and unless theBritish have built up a special resistance to it, there may be a lot worseways of working out our lottery numbers for the coming week. Justinclude a few that came up last week, or even the week before. Itwon’t increase your chances of winning, but it might (taken with ahealthy dose of salt) allow you to add the multi-prefix to your millions.

On a more realistic plane, these ideas can be applied to any number ofmarkets, ranging from the choice of dog trap at Monmore to the decisionas to whether to back the favourite in the third at Royal Ascot. Study theestablished form, be aware of the instincts of the crowd, and don’t over-react. And remember the laws of probability – the next time the housecleans up with a zero on the roulette wheel ten times in a row, it’s nomore and no less likely to come up zero next time than it is to land onyour favourite number. On the other hand, you might have a bent wheel!

85

DO MULTIPLE BETS MAKE SENSE?

Chapter 7: Do Multiple Bets Make

Sense?

You have only a vague idea how much you’ll win if your predictionsturn out to be dead on the mark, you’ll lose a lot more on averagethan by blindly backing favourites, and you simply can’t shoparound for a better price. No, I’m not talking about the NationalLottery, but the Computer Straight Forecast. The bookies absolutelylove it, which is no surprise, but so do many punters. Why, andwhat are the facts?

The CSF (or Computer Straight Forecast) is a bet designed forthose who think they can predict the first two animals past the post,in the correct order. The payout is not announced until after theevent, and is based on an esoteric computerised formula, whichincludes starting prices, number of runners, and a number of otherless obvious factors. Of course, if you are clever or fortunate enoughto solve the puzzle, you stand to earn a tidy sum, especially if on allknown form there was no earthly reason why the beasts shouldhave bothered to turn up at all.

Still, you’d have earned a rather tidier sum if you came up withthe six Lotto numbers drawn each week, but that’s not enough tomake it a good bet.

The problem with CSF bets is that unlike single bets, settled atthe starting price determined by trading on the racecourse, there isno equivalent competitive process in arriving at the computerisedodds. Thus, although bookmakers are at liberty to set odds onforecast bets, in practice the payout is almost always determined inaccordance with a national formula, which is determined by theindustry. In other words, bookmakers in the UK explicitly co-ordinate the prices of CSF bets. The potential implications of thispricing strategy are clear. Even if the prices for single bets are closeto competitive prices, the industry may build in an extra markupinto its pricing formula for CSFs.

The only real competition for this type of bet is the Exacta offeredby the Tote, which is their equivalent. Prior to the existence of theExacta the Tote offered a similar, but slightly different rival called

86

BETTING TO WIN

the Dual Forecast. This bet, like the CSF, required punters to selectthe first two horses past the post. However, unlike the CSF, the DualForecast did not require the correct finishing order to be specified.There is also one other aspect of the bookmakers’ pricing strategyin the UK that reduces the ability of substitute products to provideeffective competition for CSF bets. The pricing formula used tocalculate the CSF returns is not publicised and is difficult to obtain,although technically it is not withheld. Furthermore the likely payoutto each CSF combination is not made available to consumers duringthe course of betting. Indeed, the only information to whichconsumers have ready access is the actual CSF payout to the winningselection. It would still be possible (though not without cost) forconsumers to estimate the expected price of a CSF bet by usinghistoric data on winning payouts. However, suppliers hinder eventhis by changing the pricing formula every so often.

In summary, there are plenty of reasons why we would expectthe CSF bet to offer worse value to the punter than win bets. Is thevalue actually as bad as we would expect?

Well, as it happens, this very subject has been addressed in apaper I co-authored, with David Paton, for the journal, the Reviewof Industrial Organization.

In that paper, we compared the return to CSF bets with whatwould be expected on the basis of the starting prices of the horsesinvolved. Unsurprisingly perhaps, we found that the average CSFpayout was significantly worse, even allowing for the lower averagereturn generated by longshots. The worst value of all, we found,was concentrated in pairs of longer-priced horses.

So much for the Computer Straight Forecast, but does this meanthat all multiple bets represent poor value compared to their singleequivalents?

The way in which bookmakers frame their bets certainly suggestsso. Take the Lucky 15, for example. This comprises 15 bets, made upof all the possible combinations (singles, doubles, trebles,accumulator). The Lucky 31 bet is the five-runner equivalent of theLucky 15, while the six-runner equivalent is called the Lucky 63. Sokeen have the bookmakers been to encourage this sort of bet thatthere are a variety of special bonuses available to punters who arekeen to dabble at this end of the market.

87

DO MULTIPLE BETS MAKE SENSE?

The reason for their apparent generosity is that bookmakersassume that the margin is in their favour, and so any multiple betsimply multiplies the size of that margin. Their confidence isincreased because the punter is unable to shop around for the bestprice, but instead must take a given early price, offered bythemselves, or the starting price. So what is the advantage to thepunter of multiple bets, of this type or any other?

A clear advantage would arise if the odds on offer about each ofthe events in your multiple are genuinely in your favour. In thiscase you would be multiplying up the advantage to yourself, withthe added sugar of any free bonuses associated with the bet. Youneed to think very hard about whether this really is the case,however, especially since you are being forced to take the odds youare given by the same firm for each event in the multiple bet, insteadof having the freedom to shop around.

88

BETTING TO WIN

Chapter 8: The Real Lesson Of

Dettori Day

The year is 1996, the month is September, the date is the 28th, andthe day is Saturday. Ring any bells? Mention a certain Frankie Dettori,and Ascot racecourse, and if you follow racing (or even if you don’t)the penny may begin to drop. They certainly started dropping outof the bookies’ satchels, and into the arms of grateful punters, theday that Frankie brought home Fujiyama Crest for a historic seven-timer. An estimated 4000 million of them in fact (or £40 million).

The events of that day stand unparallelled in the history of racing,and may never be repeated. If they are, it is unlikely to involve sucha high-profile jockey, at such a high-profile racecourse, on a day asbig as a Saturday featuring the Queen Elizabeth II Stakes. Even so,there are important lessons to be learned from that day, and theyare well worth applying when considering ‘multiple’ bets.

The last race of the day provides a clue. Dettori’s mount in thelast, Fujiyama Crest, was allocated top weight, despite beingunplaced in his previous two outings. He also had seventeen otherrunners to contend with. As the firms opened for business that day,12 to 1 was freely available in the offices. Allowing for the bookies’over-round (or margin), this price implied that the odds-makersthought a bet on it would have little more chance of success thanpicking the horse’s name out of a hat containing the field.

Yet by the time that Frankie had driven Fatefully past the post towin the fifth race of the day, 4 to 1 was as good as you could get.

No sooner had the BBC’s extended coverage shown Lochangelflashing home in the sixth, than the price about Dettori’s final mountshortened even further, to 6 to 4; 6 to 4, please note, about the samehorse and jockey that had been quoted at twelves only a few hoursearlier.

Significantly, most of the liabilities on the myriad of Dettori-basedaccumulator bets had built up off-course, with the on-course layersdealing mainly in singles. The 6 to 4 about a horse with FujiyamaCrest’s claims just couldn’t last, and soon the horse was pushed outto 2 to 1. Still ridiculously short in theory, but theory has to take a

89

THE REAL LESSON OF DETTORI DAY

back seat when millions of pounds are at stake. In the event, for theprincely sum of one pound, a seven-horse Dettori accumulatoryielded, at starting price, £25000. At early prices, the same stakewould have knocked back the bookmaker almost a quarter of amillion!

For those who never did fancy Frankie’s mounts on the day, isn’tall this spilt milk? Not at all, for there is a lesson to be learned. Justlook at the details. Even by the fourth race, the odds had started totumble on Dettori’s mounts, for no other reason that that they werethe lag end of multiple bets. Very few of these had, in fact, gone thefull monty, most of them being doubles, trebles and four-horseYankees. Thus, Decorated Hero in the 3.55 had opened in themorning at 14 to 1, but was returned at half that. In the fifth, Fatefullyhad opened in the morning at 100 to 30 but was returned at 7 to 4.This tells a valuable tale, and one which in smaller measure applieseach and every day.

The point of the tale is this. If you follow particular patterns or aparticular system in placing multiple bets, beware if these patternsor your system are likely to be followed by others. For instance, youmay back all the favourites at the televised meeting, or the horses ofa top trainer, or all the odds-on shots. If so, be wary of taking theseat starting price. By the time the liabilities have mounted up, thathealthy 6 to 1 about the last leg of your Henry Cecil Newmarketaccumulator may just have turned to 3 to 1 (if you’re lucky). Theresult is that your potential winnings are halved, or worse. For suchbets, seriously consider taking the early price.

It’s a tale which the one-time odds-on offer about Marion Jonesalso tells, in an alternative but complementary fashion, in the contextof the Sydney Olympic Games. True, a very different context toracing at Ascot, but very similar market processes at work. It wasthe day when a much-hyped five-timer for Jones (Golds in 100 and200 metres, long jump and two relays) was punctured by the superiortechnique of long jump specialists Fiona May and Heike Drechsler.

Again, the lesson was the same. If you follow particular patternsor a particular system in placing multiple bets, beware if your strategyis likely to be followed by others. Whatever the context, you can besure that the bookmakers will be alert to it and will take defensiveaction. In the case of a series of horse races, the defence, as on Dettori

90

BETTING TO WIN

Day, was to shorten up the starting prices of the tail end of potentialwinning multiples. In the case of well-hyped accumulators, of theMarion Jones variety, the prices are shortened up well in advance.The price of 2 to 1 with William Hill to beat the likes of long jumpspecialists of the class of May, Drechsler and Tatyana Kotova wasmiserly, and that’s not said with the benefit of hindsight. It wasflagged up on betting advice pages across the nation. Yet still thepunters came in droves. So it matters not whether we are looking toAscot or the Millennium Games. If your accumulator bet is anobvious one, be careful before placing it. The critical question youneed to ask yourself is whether value exists at the prices availableabout each part of the multiple. Of course, you might still be gratefulfor the £25000 to a pound SP bettors won on Dettori day. Better,though, to have placed the bet early and to go on earning a goodproportion of that that in annual interest.

91

THE REAL LESSON OF DETTORI DAY

Chapter 9: Draw Biases

Every time we turn on to the World Athletics Championships or theOlympic Games we see a pattern confirmed at track-side that seemsto have held from time immemorial. The pattern is this. Sprinterswho have to negotiate a bend, whether it is the 200 metres flat orthe 400 metres hurdles, or anything in-between, dislike inside lanes1 and 2, and outside lanes 7 and 8. According to conventionalwisdom, the outside lanes give the budding champions nobody tochase, reducing their chances to that of a dog without the ‘hare’.The inside lanes, of course, give them plenty of ‘hares’ to hunt down,but those routes are apparently too cramped to bring the most outof the finely honed cardiovascular systems which are a must for themodern athlete. To make matters worse, the organisers usually seefit to allocate the fastest qualifiers the most prized berths (lanes 3 to6), leaving the poor also-rans to finish even further in their wake.

Translate humans into hounds, lanes into traps, and we just might,from a betting point of view, be on to something. Certainly, DekTerrell and Amy Farmer seem to think so, and they provide theevidence in a 1996 article in the Royal Economic Society’s flagshipacademic outlet, the Economic Journal. In a mammoth study, theyexamined several thousand races run over a period of more thanfive years at the Woodlands dog track in the USA, to see whetherbettors could secure a long-term advantage from betting particulartraps.

For academic purposes, they termed this a study into the ‘post-position anomaly’. Ordinary punters call it looking for a simplemoney-spinning system.

Common sense dictates that no such pattern should exist, for avery obvious reason. If, say, the dog in trap three regularly pays outmore and/or more often than that in trap 6, why on earth is nobodyspotting the so-called ‘anomaly’ and making a quick killing. Oncethey do, of course, the odds begin to move and the free cashpointbecomes somewhat denuded of its usual supply of notes. At least itshould, but the Terrell and Farmer study finds otherwise, and theresults provide rather interesting reading.

Instead of the returns equalising over time for all traps, the study

92

BETTING TO WIN

showed that bettors who stuck to the inside track grabbed asignificantly higher proportion of the pool than those betting eitherrandomly or on any other trap. Bets on the outside trap paid outworst.

For whatever reasons, and there may be more than one, typicalbettors at the Woodlands track consider the outside berths the bestof all, and simply bet them with excessive abandon. In other words,the dogs nearest the hare may be starting from a nice position, butcertainly not as nice as these bettors like to think. The consequenceis that far too much cash is unloaded on the outside traps, and fartoo little on number one in particular. This leaves those whosystematically back the inside track with the best deal on offer. Thisis what is called a ‘post position anomaly’.

Unfortunately, the deductions from the parimutuel (tote) systemwhich characteries the USA track more than ate away the profitsthey identified from a blind strategy of backing the inside trap oneach and every race.

Along the same lines we find a system proposed by Brian Canfield,Bruce Fauman and William Ziemba in 1987. Unlike Terrell andFarmer, they looked to the horses, asking whether trading rulesbased on a knowledge of a persistent bias in racetrack outcomescould be used to earn systematic profits. To do this, they collected asample of 3345 races run at Exhibition Park, Vancouver, Canada,between 1982 and 1984, examining win bets for the whole sample,and for a sample of longer races. The idea behind the special focuson longer races was to test their hypothesis that the greater thenumber of turns in a race, the greater would be the bias againstoutside positions.

Although they were able to identify a strategy of betting onparticular post positions, under particular track conditions, thatwould appear to offer positive profits, the size of the profits was notlarge enough to convince them that the strategy would workconsistently. Basically, they found that some post positions weresignificantly more likely to produce winners than others, but thatbettors were alert to this as well. As a result, the odds about the bestpost positions shortened by enough to make doubtful the long-termprofitability of any strategy based on such a system.

Undeterred by these findings, Sandra Betton, in a paper published

93

DRAW BIASES

in 1994, reported the results of a new investigation of post positionbias, which she based on 1,062 races at the same racetrack (ExhibitionPark) as that examined by Canfield, Fauman and Ziemba. Bettoncompared the average post position of the first three horses finishingwith the average post position to be expected if no post positionbias existed. She found the bias to be significant for the first twoplaces but not for the third.

Betton concluded that while ‘knowledge of the post positionsignificantly improves the information available from the odds …the relatively low explanatory power of these models suggests thatmore is unknown than known in the determination of racing results.’

To summarise these results, a knowledge of post position biasescan be used to improve your expected return, but there is no provenevidence that the extra advantage can in itself be turned into a sure-fire profitable betting strategy.

94

BETTING TO WIN

Chapter 10: What Goes Up, Must

Come Down!

A basketball player makes the shot, and the next and the next. Surelyhe can’t miss now! All past failures dim into the mists of time. ‘He’son a roll; he’s up for it,’ we cry. Those regular wayward efforts weare so used to seeing might have occurred in some other paralleluniverse or dream world. We always knew his shot was as sure asCupid’s arrow! Do we really think this way? An influential study,published in 1985 in the journal, Cognitive Psychology, by ProfessorsGilovich, Vallone and Tversky, reveals that we really do. Almosteveryone in the scientifically selected sample believed that playershave special streaks of good and bad performance, which defy thenormal laws of chance, as did a host of professional basketballplayers. This perception of performance is described in the academicliterature as a belief in the ‘hot hand’. The reason for this belief iswell documented. It occurs because people under-estimate just howcommonly long streaks occur by pure chance. Five heads in a rowis not an outlandish possibility, but let a player make five successfulshots in a row and we tend to believe that he’s on a roll. The vastmajority of times that the streak fails to materialise are simply notregistered in the same way. What applies to the perception of players’performances also applies to the teams they play for.

Does the ‘hot hand’ really exist? For the evidence, we must turnto a paper, published in December 1989, in the American EconomicReview. The study, by Colin Camerer of the University ofPennsylvania’s Department of Decision Sciences, is quiteunequivocal. There is no evidence of a genuine ‘hot hand’ effect inbasketball shooting, but there is plenty of evidence that bettorsbelieve in one, and that they put their money where their mouthsare in support of their belief. Particularly interesting is Camerer ’sdiscovery that bettors believe there is more permanence to losingstreaks than to winning ones. In other words, they have more faithin the ‘icy hand’ than the ‘hot’. A similar effect, by the way, hasbeen identified in the perceptions and actions of investors in thefinancial markets. A classic study by De Bondt and Thaler, published

95

WHAT GOES UP, MUST COME DOWN!

in the Journal of Finance in 1985, showed that stocks of firms thatwitnessed a recent decline tended to be available at bargain prices.The reason was that most speculators over-reacted to short-termdifficulties, and sold the poorly performing shares down tounsustainably low prices. In the end, of course, market forcesrestored the share price to a realistic level. Quite a bounce for thosewho kept a cool head, if not an icy hand, and bought at the bottomof the market.

How can sports bettors turn all this information to theiradvantage? Well, the best bet, according to these studies, is to opposefavourites that performed much better than expected in the previousweek. The reason is straightforward. The average bettor thinks thatsuch teams are at the start of a hot streak. This is a classic case ofwhat psychologists and economists call the ‘winner’s blessing’ effect.On average, these punters are just plain wrong. Still, the moneygoes down in support of their misguided beliefs, and shortens theodds of the ‘hot’ team below what they should be. As a result, theodds about their opponents are bigger than they should be. This isthe team to unload those readies into the unsuspecting satchels ofthose paid to make the punters’ lives a misery.

What applies to American sports should in principle translateacross the watery divide. Until we have bigger and better studies,we can’t be completely sure. Still, on the weight of evidence, thenext time you spot a favourite that performed just that little bit toowell last week, remember the old saying – ‘What goes up, mustcome down’.

96

BETTING TO WIN

Chapter 11: Betting In Running

Will they never learn? I speak of the compilers of ‘in-running’ golfmarkets. Let me give an example.

As the fourth day of the 2001 US Open at Southern Hills, Tulsa,drew to a close, just Mark Brooks, Retief Goosen and Stewart Cinkwere realistic contenders. Brooks, on the seventeenth, a couple ofholes ahead of his rivals, was 4 to 6 to win. His next shot found thegreen, while his challengers struck a couple of marginally waywardshots. Instantly he shortened to 1 to 5. Yet the green he had foundwas notoriously treacherous, and Brooks was the wrong end of itby no little distance. An easily predictable couple of putts later, adecent shot from Goosen and Cink apiece, and he was out to 9 to 4.Yet, in moving from 4 to 6, to 1 to 5, to 9 to 4, nothing spectacularhad occurred, little more than the random ebb and flow of shotsaround their predictable pattern.

No surprise, then, that as Cink and Goosen teed up at theeighteenth, a single shot ahead, they were granted quotes of 5 to 6each. Brooks, the so recent 1 to 5 shot, stood at 20 to 1! Without thebenefit of hindsight, was there ever better advice than backing theoutsider of three here?

The rest is history, as Brooks took it to a play-off with Goosen.OK, he lost, no doubt to the relief of those who hadn’t taken the 20to 1. But the point had been made. As it was when Bob took onTiger Woods in the 2000 USPGA Championship.

Bob? Bob who, you may ask? I refer to the 1 to 5 shot four holesout in the final round.

1 to 5 to beat the Tiger! This Bob must have been some player, orperhaps he had some lead. Well, let’s examine the facts.

The Bob in question is Bob May, the journeyman golfer rated solow in the world pecking order that he was offered at all odds up to150 to 1 at the start of the tournament. On many markets indeed hewas one of those ‘others on request’ sort of players. So perhaps theTiger was trailing by a fistful? Not at all. Just one shot back witheverything to play for.

The answer to the puzzle is much less obvious, and very much tothe advantage of those who understand how it occurred. It is simply

97

BETTING IN RUNNING

the predictable outcome of the way in which odds-makers, and inparticular golf odds-makers, price up in-running markets.

To set the scene, Tiger Woods and Bob May were more or lesssharing favouritism hole after hole, the Tiger shortening up to 1 to 2every time he produced a good shot, and back to evens every timehe displayed a less than perfect example of his game. Until, that is,he found the lush grass overlooking the putting surface. ‘Not sucha good lie!’ came the strangled cry from the commentary team. Itlooked good enough to me, and was proved so. Not before Mayhad been slashed, though, to 1 to 5 and Woods lengthened to 3 to1.As expected, a straightforward (for the Tiger at least) chip, withinputting distance of the hole, and we were back to where we startedfrom. Yet for a precious few minutes you could have taken the bestplayer in the world (and a bit more) at 3 to 1, just a shot down, tomake up the distance on the 150 to 1 rag.

A lesson learned, and it wasn’t long before it was easy to takeadvantage. The NEC invitational, played the following week, sawswings galore as Hal Sutton, Phil Mickelson, Phillip Price and JustinLeonard fought it out for the honour of finishing second to the Tiger.By following any of these players, even half a course out, and simplywaiting for a wayward shot, you could see your odds double ormore, before eventually bouncing back to equilibrium as professionalgolfers showed the odds-setters the way in which to play a recoveryshot. Each player in turn became hot favourite (even odds-onfavourite), with scarcely a change on the scoreboard. It was as if theodds-setters were calculating the effect of a dodgy drive on the basisthat they themselves were having to play the wayward ball.

This method of setting odds is a little strange, to say the least, andpotentially very profitable for the clued-up late night viewer withteletext buttons at hand. Strange yes, but quite in keeping with thescience of betting and indeed the theory of finance. Small changeshave a tendency to produce big effects, in the very short run atleast, until equilibrium is restored, as the overall impact of marketforces re-establishes order and good sense.

How can these professional odds-makers be getting it so wrong?Well, maybe they aren’t from their point of view. It could simply bean appropriate defensive response to the expected response of thebetting public. After all, if a player produces a bad shot, those

98

BETTING TO WIN

wavering at the margin who pick up their telephones are more thanlikely doing so to oppose him.

Now is your opportunity. Now is the time to side with the newunderdog. In truth, you may never get a better chance.

99

BETTING IN RUNNING

Chapter 12: The Difference Between

Risk And Uncertainty

To illustrate the difference between risk and uncertainty we needlook no further than the game of tennis, and in particular the 2001Australian Open semi-final between all-American Andre Agassi andAussie hero, Pat Rafter.

At the start of the match, 8 to 15 was the best you could get aboutthe American, with 7 to 4 available about the local boy. The odds towin the tournament were 5 to 6 Agassi, 5 to 2 Rafter.

After two hard-fought sets, Rafter seemed to have the advantage, losingthe first 7 – 5, but winning the second 6 – 2. Still, Blue Square, betting inrunning, gave the pre-game favourite the edge. After three, Rafter led bytwo sets to one, thanks to a dominant tie-break performance, and bynow the market-makers at Blue Square had to acknowledge that thebalance of risk and (quite literally) return had tilted. More than simplyacknowledge the fact, they went 1 to 3 the Aussie.

All conventional stuff so far, until the bearded antipodean seemed,at least to perceptive viewers, to take on a somewhat stilted gait, notto say a limp.

Risk had now turned into what students of probability theorywould term uncertainty. The difference between the two, while notobvious in normal conversation, is critical to the statistician. Whereasrisk is measurable, at least in principle, uncertainty is not.

To clarify, if I draw a card from an unopened pack, I am taking arisk. I know, if it is a fair pack, that my chances of drawing a spadeare 1 in 4. Now, if I am told that there will be a card game, but thatthe nature of the game and the organisers of the game are yet to bedecided, I am entering into the realm of uncertainty. Before, I wasable to attach a statistical probability to the outcome. Now, there areso many imponderables that it is unclear how I should go aboutestimating the odds of winning.

To return to the tennis, we can calculate from past evidence howoften a slightly lower-ranked player, two sets to one up, goes on towin the match. If I know the odds better than the market-maker, Ihave the edge.

100

BETTING TO WIN

A hobbling tennis player, two sets to one up, is a whole differentproposition.

The behaviour of the market-makers in these circumstances wasvery instructive, for it shows how professional odds-makers reactwhen the odds are not readily calculable. They suspend the betting,a stance which continued even when the afflicted Australian wasable to win his first service game of the final set.

It is interesting, therefore, that professional odds-makers are morethan happy to offer odds about a wide category of events which areequally uncertain in nature, in the sense that there is no clear wayto assign a probability rating to them.

How do they achieve this? Well, one obvious ploy is simply tosidestep genuine uncertainty, and instead offer odds only wherethe apparent uncertainty is in actual fact a case of scarcely concealedcertainty – that it won’t happen. Prince William to wed BritneySpears. Italy to win the Rugby Grand Slam. Elvis to pilot an alienspacecraft into the Loch Ness Monster. All are faintly possible, Isuppose, but you get the idea.

Such bets need not detain the sophisticated bettor for long, then,which is why the true sleight of hand is in offering odds about eventsthat rate somewhere in-between risk and uncertainty.

If you really do want to back a longshot, enter a market wherethe odds at least bear some semblance to the chance of it happening.That normally means conventional markets, subject to the cuttingedge of competitive market forces.

Having said that, we must not forget a certain Mr. David Threlfall,for it was he who once asked what price William Hill would givehim about a man walking on the moon by 1970. He was astonishedby the answer. 1000 to1! Astonished because he had just seenPresident Kennedy promising that very thing on the TV.

On the evening of 20 July, 1969, while Neil Armstrong and BuzzAldrin sat in the Eagle at Tranquillity Base, Mr. Threlfall was presentedwith his cheque for £10,000 live on the David Frost show.

David Frost is still around, so are William Hill. Similar naiveté byodds-makers, unfortunately, is in much shorter supply. The bestyou will now get is a 500 to 1 quote that the US President will declarethat it never really happened. Hardly appealing, but then again!

Such long-term predictions are growing increasingly popular, and

101

THE DIFFERENCE BETWEEN RISK AND UNCERTAINTY

make traditional ante-post betting look rather tame. Take the well-known bookmaker who is offering odds about the year that ‘man’will first walk on Mars. Still, I think I’ll leave alone their 50 to 1 aboutit happening this year.

William Hill were even more far-sighted, offering 5 to 1 at onetime that Senator Hillary Clinton would become President HillaryClinton before she shuffles off this mortal coil. Betting to close on 11November, 2040. Do they know something we don’t?

Ideally though ‘super-ante post’ bets should have a payout datethis side of an age at which you are still likely to care. Any takers onthe date of the end of the world?

102

BETTING TO WIN

Chapter 13: Strike While The Iron’s

Hot

If you leave a drunk in the middle of a field, where are you mostlikely to find him? Not a riddle, but the subject of seriouscorrespondence in the scientific journal, Nature, in the year 1905.The answer the scientists came up with is simple enough. You aremost likely to find him where you left him. This is not because hewas too paralytic to move at all, but because he is equally likely tohave moved in any direction. Since you can’t tell which, your bestguess as to his current location is where he started out. Put thisreasoning into a mathematical framework, develop all itsimplications, and you have before you what statisticians call a‘random walk’. All sorts of naturally occurring phenomena follow aso-called ‘random walk’, the best documented being pollen seedsin a liquid solution. It is, in fact, absolutely impossible to predict thenext movement of these seeds from their previous movements. Thepath is completely random. Scientists, by the way, call this haphazardjourney a case of ‘Brownian motion’, after nineteenth century-botanist Robert Brown who first spotted its existence. Random walktheory, however, has much more important implications than beingable to find that the friend you left in a field slightly the worse forwear, or that elusive pollen seed. It tells us about patterns ofmovements in all sorts of vital areas, and that includes financial andbetting markets.

The idea of a random walk in the movements of share prices hasbeen with us since Louis Bachelier first noticed that you couldn’tpredict the movements of futures prices on the Bourse (the ParisStock Exchange) from previous movements. The reason was quitesimple. At any point in time, share prices should reflect all availableinformation. At least, that’s the theory. If so, there is no way to predicthow the share price will move again, unless and until newinformation becomes available. Any movements before then will bepurely random, and as such unpredictable. As soon as newinformation does become available, the gain goes to those who spotit first. If it’s good news they should buy the share now before the

103

STRIKE WHILE THE IRON’S HOT

price goes up; if it’s bad news they should likewise sprout wings tosell, before they sit facing a loss. Unfortunately for the vast majority,by the time they have acted it’s too late. The fleetest of foot, with thequickest access to new information, have already reaped the spoils.

The same applies to betting markets. By the time the market hassettled down, the real gains to be made have for the most partdisappeared into the mists of time. The on-course racetrack bettingmarket is a clear case in point. The market may be open for severalminutes before a race, but by the time it closes the prices availablewill more than likely incorporate all the information that is availableabout the horses or dogs springing into action. That’s a pity forthose who always take the starting price, but the beauty of the Britishsystem is that you don’t have to wait. If you have genuine newinformation, it is vital to strike before the information filters out tothe madding crowd. Fortunately, because you are able to take a price,what happens after you have plunged in doesn’t matter a jot. That’sall very well if the information you obtain is the genuine article, andit’s why in those circumstances the bookies’ telephone lines arejammed with those all at once seeking to get on before the pricegoes. If, on the other hand, your new information is simply a rumouror a whisper from an unreliable source, take the advice of numerousstudies. On average, the starting price is the best price ever available.On average, of course, you won’t win at starting price. That’s whygenuine information, acted upon very quickly, is a recipe for long-term betting success. At least that’s what science tells us, and it’sbeen doing so ever since it was used to track down that drunk in afield.

104

BETTING TO WIN

Chapter 14: Prescribing The Right

Dosage

A flying visit to Louisville, Kentucky, followed by an even more hecticschedule based around the financial downtown district of SanFrancisco may sound a bit like fun. And for me, it was. But it wasalso a lot of hard work, attending and presenting papers at oneconference devoted exclusively to racetrack betting (at the Universityof Louisville), followed by a series of sessions on betting analysismore generally (the 76th Annual Conference of the WesternEconomics Association), in San Francisco.

The participants at Louisville hailed from every corner of the globe,each seeking to introduce the latest research techniques to examinewhat seemed almost every angle of betting and gaming.

Some had made their research pay – handsomely; some were init for the sheer pursuit of academic excellence; all were fascinatedby the workings of betting markets; and most turned a great dayout at Churchill Downs into an even better one courtesy of the USTote, the parimutuel.

Even the traditionalist Louisville Courier saw fit to devote a wholepage to the proceedings, and in particular to the latest developmentsin the science of forecasting which horse is likely to finish first pastthe post.

Of all the new research efforts that week in May, 2001, I shallfocus here on the offering of the famed guru of the track, ProfessorWilliam (‘Dr. Z’) Ziemba.

Dr. Z, as he is generally known, pinpoints a technique for selectingthe winner of defined high-class races, based on blending two‘experts’: the odds and the Dosage (a system based on breeding). TheZiemba theory is that bettors should choose a horse that is favouredby bettors (but isn’t the favourite) and also has a low Dosage Index.

The Dosage Index, loved and loathed by American racing fans inalmost equal measure, was invented by racing expert, Steve Roman,and popularised by the USA racegoer’s bible, the Daily Racing Form.It is loosely based on the work of a French cavalry officer who advisedthe Aga Khan on his breeding operations in the early 1900s.

105

PRESCRIBING THE RIGHT DOSAGE

Roman came up with a numerical formula based on breeding,which predicts whether a horse is capable of winning at a distanceof 1 mile 2 furlongs or further in the spring of its three-year-oldseason.

The lower the number, the better the prospects. Only threeKentucky Derby winners, for example, have had a Dosage rating ofgreater than 4.

The reason that Dosage is so critical to these types of races is thatgenerally the horses have never before run at these distances. Inother words, you don’t know if they’ll run out of steam before theend of the race. As Dr. Z puts it, ‘The hardest race of their life is theDerby … and they’ve never gone that distance.’

Moreover, the beauty of the system is that bettors don’t have toanalyse complex mathematical equations, or even know much aboutthe horses for that matter.

To put it simply, the Dr. Z system for predicting the outcome ofClassic races run over a mile and a quarter or more is to exclude thefavourite, and of the remainder to back whichever horse has thebest stamina pedigree.

The devil may lie in the detail, but for a theory that has producedsuch consistently healthy profits over such an extended period oftime it really is almost as simple as it gets.

106

BETTING TO WIN

Chapter 15: The Home Ground Factor

I wonder how many people will still be leafing through a reprint ofthis book in 2020. I hope quite a few. In any case, judging from whatI found recently in a faded 1982 edition of New Scientist, you neverknow what you can learn from the past.

The 1982 article, entitled, ‘Why Spain should win the World Cup’,set out to answer whether it was England that won the World Cupin 1966, or whether it was Alf Ramsay and the boys. In other words,how much did home advantage shape the outcome?

It is conventional wisdom in 2001 that home advantage confersabout half a goal superiority compared to playing on neutral territory,so that playing on home territory confers about a one goal advantagecompared to the reverse fixture.

Jack Dowie, who has studied betting markets for longer than most,knew nothing of 2001 when he sought to explain in the New Scientistwhy it is that professional sportsmen should be so cursed by theirtravels. He knew rather a lot, however, about football statisticsbetween the fall of Hitler and the then recent decision to awardthree points, instead of two, for a League win. Over this period hefound that home advantage was worth about 0.6 of a goal for teamsplaying in the top division, and about 0.7 for the lower divisions.That means that the value of home advantage has weakened slightlysince the early 1980s, and particularly so in the lower echelons ofprofessional football.

Dowie’s article sought to distinguish between the effects of threewell-established explanations of home advantage, known as theThree Fs explanations. Each of these Fs – fatigue, familiarity and fans,has attracted the support of academics over the years as the keyfactor explaining the under-performance of the travelling set.

First, fatigue. The 1982 study looked for evidence that away teams’performances drop off relative to home teams as the gameprogresses. In his sample of nearly 40 years of data, Dowie foundlittle evidence of any such effect, as measured by the likelihood ofscoring a goal at any given point during the course of the match.Away teams did score fewer goals, on average, than home teams,but this disparity got no worse as the game developed. Clearly this

107

THE HOME GROUND FACTOR

evidence does nothing to strengthen the case of those who point tothe weariness of the visitors as a key issue.

Likewise, familiarity with the pitch is likely to be a bonus for thehome team. If this is a key factor, however, teams who are travellingfrom a similar pitch to the home team should be less disadvantagedthan those who are travelling to a very different sort of pitch. Oneobvious way to test this is to ask whether teams who play onrelatively big pitches have a particular statistical advantage whenplaying host to visitors whose own home ground boasts a smallpitch, and vice versa. In fact, there appears to be no such pattern inthe more than ample data set. Home advantage seemed to remainconstant whatever the relative pitch sizes of hosts and visitors. Notmuch support there for the familiarity hypothesis.

That leaves the third F – the fans. They obviously count forsomething, but inasmuch as they do, what is it about them? Is it theabsolute number of fans, or is it the relative number of home fanscompared to away fans? Here the data set did produce someinteresting findings, revealing that the advantage conferred byplaying at home was significantly greater for games played in thelower divisions than in the top division, even though the absolutenumber of supporters was much smaller in these games. Moreover,the advantage was much less in ‘local derbies’. The conclusion to bederived from all this is that it is the balance of support at the groundwhich matters.

A study, quoted in The Times (9 May 2002) may provide part of theexplanation for this crowd effect. A study by Alan Nevill, of theUniversity of Wolverhampton, set out to test whether referees’decisions are affected by the noise of the crowd.

In an experiment, 40 qualified referees were shown video footageof 47 tackles from a Premiership match. The referees were dividedinto two groups, half of whom were exposed to the originalsoundtrack, while the other half listened to a silent version of thematch. Neither group had access to the original referee’s decision.In actual matches, it should be noted, about 60% of bookings points(10 for a yellow, 25 for a red) are awarded to the visiting team.

Those referees who watched the original soundtrack werereluctant to penalise the home team, judging 15% fewer of the tacklesby home players to be fouls as compared to those referees who

108

BETTING TO WIN

watched the silent footage. The implication of the study was that inthe absence of crowd noise the officials were more equal-handedbetween the home and away sides. The original referees’ decisions,however, more accurately mirrored the behaviour of those armchairreferees who had access to sound. Alan Nevill has his ownexplanation: ‘To get the crowd off their back they wave play on.’

If true, this explanation of home advantage is fascinating, but it isonly of use to a professional gambling strategy if it is not alreadyfully factored into the bookies’ odds. Insofar as it is not, bettors whoback teams where the ratio of home to away fans is particularlyhigh will, in the long run, be on to a winning strategy.

If it were only so easy, we would all be millionaires, of course. Onthe other hand, perhaps we should be.

109

THE HOME GROUND FACTOR

Chapter 16: The Patriot Bias

It was a great day to be at Trent Bridge, and especially to be seateddirectly behind Merv ‘the Swerve’ Hughes, ‘Pistol’ Paul Reiffel andhalf the Aussie throng. Indeed, for a brief few hours on thatmarvellous afternoon in the summer of 2001 I thought that we wereabout to witness a reflection of the glory days of Headingley 1981.As Aussie wickets were scuttled in quick succession, I was all themore delighted to be entertaining that very evening a visiting cricket-mad friend hailing from down under.

It was not long before our evening meal turned to talk of wagering.‘Bet you’re glad you took a slice of that 11 to 1 about an Englandwin?’ he offered, knowing all too well that I had done no such thing.Shamefacedly, I had to admit I had not even taken the 14 to 1 availableelsewhere. Once again, it seemed, I had fallen into what I have myselfall too often called the ‘patriot bias’, which is the tendency we Britshave to swing from crazy overconfidence whenever our team isfavourite to win, to crazy underconfidence when we are the haplessunderdogs. By next morning, I was all the more aware of thissyndrome, as I had to call on every dispassionate bone and sense ofreason to keep the ‘phone on the hook despite that ‘massive’ 13 to 8about England routing the Aussies.

The cold logic of the fact was that both odds were wrong. 14 to 1was too long at the start of a match where the draw was always avanishing possibility, and where so much could turn on theresolution of a few crunch moments which could go either way. 13to 8 was really too short to take about England on the basis of a late-afternoon rally in the failing light.

The bottom line is that we as a nation of sports lovers are all tooready to take ridiculously short odds in any sport where our nationalteam is expected to win, and to pass up ridiculously long oddswhenever we as a nation are expected to lose.

The overconfidence side of the equation has been given amplesupport in a 1995 US study of so-called ‘teaser ’ bets, undertaken byJoseph Golec and Maurry Tamarkin. In teaser bets, the bookmakerallows the bettor to be wrong by a given amount and still win, albeitat the expense of a lower payout. It appears, according to the Golec

110

BETTING TO WIN

and Tamarkin study, that bettors in fact underestimate dreadfullyjust how likely they are to get things wrong, and so believe that alittle leeway is much more useful to them than, in fact, it actually is.In other words, bettors are too confident in their own abilities.

Apply this now to the national team playing football or cricket, orwhatever, and we seem to take this overconfidence to an extreme. Ifwe think England are going to beat the USA at football, say, we area little too sure of ourselves, and the odds will be too short inconsequence. If we think that England are going to lose to the touringside, on the other hand, we are too sure of the visitors, and theodds about England will tend to be too long.

Be careful, because the ‘patriot bias’ is still only a theory. Even so,it is a theory in which I have sufficient confidence that I shall bebacking England to win the next Ashes series held on our shores.

Of course, I may lose, but the ‘patriot bias’ tells me that at least Iwill have grabbed the value. Now what was that I said about beingoverconfident?

111

DRAW BIASES

Chapter 17: Small Is Beautiful

Heads or tails, Greg Rusedski or Tim Henman? What do thesechoices have in common? Well, it’s a choice between two, of course,and there’s going to be a result. As a bookmaker, what odds do youset? Assuming the coin has two sides of the conventional variety,fair odds are evens the head and evens the tail. This means thatwhatever sum of money you stake on the outcome will be matchedby the bookmaker if you win. You’ll also get your stake back. If youcall heads, however, on the coin that landed on its tail, you losewhatever sum you bet. At such odds, over the long run you (andthe bookmaker) will break even. That’s of little use to the bookiewhose motivation is to make a living, so it’s more likely you’ll beoffered, say, 4 to 5 each of the two. These odds means that for everyfive pounds staked you can expect to win four. Worse still, thereseems to be no way, short of fixing the coin, to beat the odds. Not sowith the tennis match, where your perceptions of the true oddsmight well differ from those offered by the bookie. You may or maynot be on the right side of the difference, but at least there’s hope,and you may even be able to shop around for different odds withdifferent companies.

There’s something else as well which makes such bets particularlygood value. The notional ‘profit margin’ built into choices betweentwo outcomes is normally particularly low. The term used to calculatethis margin is known as the ‘over-round’, and it works like this. Ifthe bookmaker offers you evens about Rusedski and evens aboutHenman, you can stake five pounds on each. Now it doesn’t matterwho wins the match, you will win five pounds and have your stakeon the winning outcome returned to you. So by covering all theoptions you get ten pounds back for the ten pounds staked everytime. The ‘over-round’ in this case is zero. If, however, you are onlyoffered 4 to 5 about Rusedski and evens about Henman, you can’tfor sure cover your bet. Stake five pounds on each of the tennisstars, and you will earn five pounds if Henman wins (£10 includingyour stake back), but only an additional four (£9 including yourstake back) if victory goes to his opponent. In effect, you have stakedten pounds to be sure of receiving back nine. The percentage

112

BETTING TO WIN

difference (usually added to 100%) is known as the ‘over-round,’and the smaller the over-round the less in-built advantage thebookmaker has in the odds.

Indeed, in events like tennis, or even boxing (where the draw isusually quoted at 25 to 1 or longer) a simple calculation of the over-round implied in the odds reveals just how disadvantageous theyare to the odds-makers, relatively at least. The reason may be linkedto the sheer simplicity of the outcome. Either A wins or B wins.That doesn’t leave a lot behind which the bookmaker can hide, andone of the consequences is a relatively small ‘margin’ on such betsfor the bookmaker.

This is why bookmakers promote heavily the expansion of choicesthat arises from choosing the exact round of victory in boxing, orthe precise score in tennis, or the number of frames in snooker.

To maximise your advantage, avoid the pre-tournament odds.There are so many options that the bookies can shade the prices fartoo easily. This doesn’t apply to individual match bets. Take as anexample an imaginary Wimbledon men’s singles final. Let’s say thatLleyton Hewitt is available at 8 to 15, while 7 to 4 is the offer aboutAndre Agassi. At these odds it is almost possible to bet on bothplayers, to suitable stakes, and win whatever the outcome. Not quite,but the ‘margin’ in favour of the bookies works out at a mere 1.6 percent. In a number of cases, however, the margin disappearsaltogether – say Venus Williams at 6 to 5 with one firm, playingSerena at 5 to 6 with another. Now compare this with the racetrack,where the notional ‘margin’ in favour of the bookmaker, at startingprices, works out at about 2% per horse in the field. In other words,a 10-horse race will probably be characterised by a ‘margin’ of about20%, whereas it might be nearer 50% or even 60% for a largesponsored handicap.

As a bookmaker, any means of increasing the number of optionsis usually a good thing. This can mean publicising the ante-postodds about a championship very heavily, or the odds at the start ofthe tournament, as in both cases all the players are still in the field.The other method is to encourage betting on the score, which offersfour options for the women’s game (2 sets to 1 and to love in favourof each player) and up to six for the men’s game (3 – 0, 3 – 1, 3 – 2 forboth). The ideal scenario, from the bookmaker’s point of view, may

113

SMALL IS BEAUTIFUL

be to offer odds at the start of the tournament about the number ofsets an identified player will take to win the final.

The best advice from a betting point of view, therefore, is to watchand wait for individual match odds. From the quarter-finals on, thenumber of bookmakers offering odds start to grow, and the overall‘margin’ in favour of the bookies (if indeed it is overall in their favour)starts to plummet.

Now is the time to step in, as soon as the odds are generallyavailable with most bookmakers. This means, for example, lookingto bet on the Sunday men’s final on Saturday morning and thewomen’s final on the Friday. By the day of the match the mostgenerous odds will have been taken.

Indeed, excluding the 40 to 1 no-hopers and beyond, the over-round, at the best odds available with leading bookmakers, issometimes well under 100%, which means a healthy margin in favourof the betting public. This means that if you are quick enough offthe mark, it is often possible to stake at a variety of odds such thatyou are guaranteed a net profit before another ball crosses the net!Of course, we expect lower margins when field sizes are smaller,but this is a gift from the tennis gods which never ceases to amaze.

Much of it has to do with wide disparities in the judgements oftennis odds-compilers. Take the 2000 Australian Open, for example.At the quarter-final stage, while Jennifer Capriati was 5 to 1 withCoral, she was 7 to 2 elsewhere. While Serena Williams was 7 to 1with Chandler, she was 9 to 2 with Ladbrokes. While Justine Heninwas 8 to 1 with William Hills, she was 9 to 2 with Ladbrokes. WhileKim Clijsters was 12 to 1 with Coral, she was 6 to 1 with Chandler.

This is just to look at the main market-makers. Those who havethe time and energy to peruse the Internet for the widest range ofoffers would have spotted even larger disparities.

The odds-setters can’t all be right, of course, but where tennis isconcerned they manage to stay in business. It is difficult to fathomlogically why this is so, but part of it might lie in the money that thebookmakers make on so-called special markets. Try not to getsidetracked into these, however cleverly designed, for they arespecial in one very particular sense – in the degree to which theodds are tilted towards the book.

For the record, I was not tempted by Blue Square’s one-time offer

114

BETTING TO WIN

of 25 to 1 about Anna Kournikova kissing (I think it was) the ball boy!So choose the player who you think offers the best value at the

quarter-final stage, and stick to a bet on that player to win thetournament. The value in such a strategy is particularly marked inthe women’s game, where the odds tend to diverge more sharplyacross bookmakers than is the case with the men’s game. Why thisis so is anybody’s guess, but it matters not to the profit-seekingpunter.

So, with a major tournament approaching, how to go aboutidentifying this value? The first thing, once the quarter-finalists areidentified, is to strike off any player offered at 7 to 1 or above by themost miserly bookmaker. This bias against longshots seems to beexceptionally strong in women’s tennis.

Of the remaining handful, simply choose the player whose longestodds bear the greatest ratio to their shortest odds. This may not bethe winner, but to value-seekers it probably represents the bestavailable value. And for those who must bet early, you could doworse than stick with the obvious.

Finally, if you really do favour the racetrack, be keenly aware ofone fact. Horses starting at short odds are just as good value in alarge field as a small field. Thus, a 1 to 2 shot has basically the samechance of winning if up against twenty rivals as up against one, i.e.a very good chance. Remember this when selecting value.

In conclusion, it is regrettable that the impact of high tax rates onwinnings has historically discouraged even clued-up bettors fromexploiting the value available in small fields of whatever complexion.With the new era of tax-free betting, however, such excuses arebecoming increasingly outdated. Small really is beautiful at theracetrack and in so many other sporting arenas. The value is there –grab it with both hands!

115

SMALL IS BEAUTIFUL

Chapter 18: The Hi-Tech Crystal Ball

Wouldn’t it be wonderful if we could devise a computer programthat could instantly tell us the winner of any race? What would wegive for such a hi-tech crystal ball? The answer depends, of course,on who else has access to this magic oracle. If it was freely availableto all and sundry, the information would, of course, be useless. Justimagine, however, that you and maybe a select group of other like-minded souls had the secret. Now we’re talking. Just stop dreamingfor a moment, and let’s take a trip into the real world. In the land ofthe waking it is, of course, impossible to predict the winner of everyrace, if only because there are too many random factors that canupset the best of calculations. Statisticians call these random factors‘noise’. That’s the bad news. The good news is that these factorstend to balance out over time, so any properly devised system whichtakes into account all or most of the predictable indicators will tendto do very nicely indeed. The question is whether such a systemexists, or indeed whether it has ever existed. Ask Bill Benter how hemade his millions, and you have your answer.

Bill Benter is a name that is most famously associated with thebetting syndicate that cleaned up in the Tote-style pools characteristicof the Hong Kong racetrack. Not only a successful gambler, he hasalso contributed to the academic literature, notably in a paperpublished in 1994, entitled ‘Computer based horse race handicappingand wagering systems: a report’. The question he seeks to answerin that paper is whether it is ever possible for a fully mechanicalsystem (a system based on a fixed model) to beat the races. Ever thepractitioner, he provides the answer by constructing just such amodel. The method is to identify each individual factor that couldpossibly predict the outcome of a race, and then to whittle thesedown to the most reliable and effective. Once he had a model thatworked on past data, he then tested it practically on a large sampleof further races. This so-called ‘out-of-sample’ testing is vital for aproper statistical analysis. Sometimes he found that a variable wasuseful in predicting race outcomes, but he couldn’t really understandwhy that should be the case. In such circumstances, he decided,the best policy was not to care. Faced with a choice between a

116

BETTING TO WIN

profitable model that he couldn’t fully explain, and an unprofitableone that he understood perfectly, he chose the former. If it works,as the saying goes, don’t fix it.

So what’s the bottom line? Well, that’s the best part. Benter ’s modelwas tested over five seasons, and consisted of around 470 bets ayear. Although the average track take was about 19%, the modelshowed a significant net profit in four of the five seasons examined,and a sharp upward trend in the returns over time as the modelwas improved.

Fortunately for Bill Benter and his team, they got in early, whenthere was little competition from other computer wizards. It was, inhis own words, a Golden Age for such a system. As he so clearlyobserves:

‘In the future computer handicappers may become morenumerous … which will likely cause the market to becomeefficient to such predictions. The profits have gone, and willgo, to those who are “in action” first with sophisticated models.’

Unless, of course, you have access to that model which is just a littlebit better than the others.

117

DRAW BIASES

Chapter 19: Better Late Than Ever!

It’s the last race of the day, and your wallet is weighing rather lessthan it was in those expectant hours preceding your trip to the races.You’ve just enough left for a pie and a pint – and one last tennerover. You have two choices. You can make a quick exit, or you canstay and fight back. Now, assuming you decide to be bold, just howbold a strategy should you adopt? Fortunately, research undertakenin the USA and the UK may have the answer.

As long ago as 1956, an article appearing in the American Journal ofPsychology offered the first clue. The author, William McGlothlin ofthe Rand Corporation, found that while betting on favourites was amuch better strategy overall than betting on longshots, betting onfavourites in the last race of the day was even better. The evidencein the psychology literature was soon backed up by professionaleconomic research, Mukhtar Ali of the University of Kentucky (1977)confirming the findings for a very large data set of races run in theUSA. Indeed, the evidence was quickly so overwhelming that it wassoon accepted as fact, and given the status of a ‘Law’, arguablyputting it rather perilously on a par with Newton’s Laws of Physics.The ‘Law’ is known to academics as Gluck’s Second Law, and itstates that ‘The best time to bet favourites is in the last race.’

Why should this be so? Explanations abound, but the simplestexplanation may well be the best. It goes like this. As the day wearson, the average punter at the track has less and less money to playwith. By the time of the last race, there may be less available to stake,but there’s still the hope of at least breaking even on the day. Butonly if you bet the longshots. That tenner on the 20 to 1 shot willsend you home as happy as Riley on a good day. A tenner on theodds-on good thing, on the other hand, might reduce your losses ifit wins, but you’re hardly likely to depart the track with a broadgrin on your face. That’s why the last race of the day is often soaptly named the ‘Getting-out Stakes’.

As a shrewd bettor, you can learn from this pattern of behaviour,to your long-term advantage. If the mug punters are betting toomuch on the longshots, and doing so in spades as the sun begins todip ever lower in the sky, that must mean that there’s relatively less

118

BETTING TO WIN

going on the shorter-priced nags. The consequence is that thefavourites are better value than they ever were. If you’re going tostrike at all, this is the time!

Surprisingly perhaps, a study of the behaviour of off-course bettorsin the UK, conducted by Johnnie Johnson and Alistair Bruce, andpublished in 1993 in Psychological Reports, showed that betting shopregulars tended if anything to bet the favourites later in the daymore heavily than ever.

For off-course bettors, of course, there’s always another day,another pound to win or lose, so the same psychology as afflicts theirregular racegoer doesn’t necessarily apply. Alternatively, it couldbe that the British are more sanguine about the cash that’s alreadyflown the nest, and are not half as impatient as their Americancousins in trying to grab it back. Don’t bet on it, though.

So what’s the bottom line? Should we hang around for the lastrace of every meeting before unloading our previously zipped walletsonto the latest sure thing? The bottom line is in fact to heed theevidence, but to use it with care. There is still not enough researchdone in the UK or on recent data to provide convincing evidencefor a system. Moreover, before you decide to hand over thick bundleson the favourites in the last race of every meeting, please note thatthe strategy at its best doesn’t predict a clear profit to a blind policyof betting every favourite in every last race of the meeting. It simplypredicts a better return than you could otherwise expect.

119

DRAW BIASES

Chapter 20: The Value Of Tipping

Agencies

Compared to other areas of betting research, there has been verylimited academic study of the value of professional racetrackforecasting services. Indeed, the most well-known study, at leastfor the British racetrack, until fairly recently, was published byNicholas Crafts in 1994.

Crafts examined the performance of three betting systems on salein the market. He tested each of the systems on the basis of resultsgenerated prior to and after publication of the systems. He foundthat one of the systems demonstrated some evidence of pre-taxprofitability prior to publication of the system, but otherwise noevidence at all of any significant profitability.

He also reported the absence of profitability in nine racing systemsand five tipping services, based on results reported elsewhere andconcluded: ‘The continued sale of these systems suggests that theparticipants in British horserace betting include many gullibleoutsiders.’ By ‘outsiders’ he was referring to punters without accessto inside information.

I decided to put this conclusion to the test using tips provided byfive relatively expensive and well-publicised tipping services in the1990s. I have kept the names of the services anonymous, butpublished my findings in the highly respected academic journal,the Journal of Forecasting, in November 2000.

What I found was perhaps surprising. All of the services examinedgenerated a pre-tax profit to the advised staking and betting plans,at starting price. In no case, however, were these profits large enoughto be judged significant using standard statistical tests.

It was also possible in some cases to improve the profitability byfocusing only on the more strongly advised tips.

It is quite feasible that the results would have been better at theprice available when the advice was first announced. This is becausethe odds may subsequently have shortened, in response to moneyplaced on them by clients of these agencies.

My overall findings, therefore, did not show evidence of

120

BETTING TO WIN

spectacular success, at least at starting prices, but it was certainly amuch happier tale than the one told by Professor Crafts. Even so, inassessing these findings, there are still certain problems to beaddressed: The problems are these:

1. Is it possible to obtain the advised price? If a large numberof clients are all vying for the same price, it is quiteconceivable that the price would be shortened before asignificant number could obtain that price. Indeed, thedefensive response of a bookmaker or bookmakers targetedin this way may be to shorten the price excessively, evenbelow the subsequent starting price.

2. How much is it possible to stake at the advised price? Nobookmaker can be forced to stand the quoted price to anygiven sum. It is well known that bookmakers, especially inthe face of a heavily tipped horse, will restrict the stake atthat price.

3. How much does the tipping service cost? The more expensivethe service, the greater the stakes required to show a net profit,even if the tips produce a theoretical profit. Of course, thelarger the stake required, the greater the risk involved, andany returns must be weighed against this additional risk.

4. Be aware of how any profits advertised by the forecastingservice are compiled.

Blatant dishonesty is one thing, but there are more subtle methodsof generating a notional profit. Let me point out one of the moresophisticated ploys.

Take a service which works on the following basis:The client is provided with a tip about a horse, along with a price

(a so-called ‘minimum price stipulation’).The client is told to bet on the horse only if the minimum price

becomes available. If it does not become available, the advice is void.Say, for example, that the tip for the day is Mr. Nippy, with a

minimum price stipulation of 4 to 1.

121

THE VALUE OF TIPPING AGENCIES

Now assume that Mr. Nippy opens in the market at 3 to 1. On thebasis of the minimum price stipulation, you should wait, for this isbelow the minimum price.

Say, now, Mr. Nippy lengthens to 4 to 1 in the market. What shouldyou do? Take the 4 to 1? Clearly you should, because you do notknow if it is going to shorten or lengthen thereafter.

For example, say it subsequently shortens to 3 to 1 and you havenot taken your chance to bet at 4 to 1. It is now too late.

The price may, however, lengthen to 5 to 1. What now? You havealready placed your stake at 4 to 1, in case it shortened again.

The tipping service in these circumstances has the best of bothworlds in announcing its results.

If the price extends to 5 to 1, or 6 to 1, or even longer, they canclaim the longest price it ever reached. If it shortens after touchingthe minimum price stipulation, they can claim the price before itshortened.

Hindsight is a wonderful thing, of course. It is available incompiling their notional profit and loss accounts, for display to thepublic, but to the hapless client no such perfect foresight is available.

In other words, the accounting option open to the service (withoutlying at all) is not available to the client, and significant profits canin this way potentially conceal something far less flattering. Clever,isn’t it?

This is not to imply that every tipping service that works on thebasis of a minimum price stipulation necessarily operates like this.Before parting with your cash, however, it may repay your time tocheck it out.

In conclusion, there is plenty of evidence, from the analysis offorecasting services highlighted here, and from studies of theexistence of inside and superior information examined elsewhere,to suggest that some people know enough to make substantial profitsat prices obtaining at some point in the market.

But remember the old adage – those who are telling don’t know,and those who know aren’t telling. Be careful, therefore, to checkout the credentials of the service, and to factor in all the costs.

122

BETTING TO WIN

Chapter 21: Betting On The Oscars

Not since the days of Pretty Woman had Julia Roberts been availableat such a generous price, although even Richard Gere would haveneeded a betting brain to take advantage of the offer.

I refer here to the 11 to 10 available at best price in my sampleof bookmakers at the opening of the market about Ms Robertstaking the 2001 Best Actress Oscar for her part in Erin Brokovich.Although available only to a maximum stake of £25, it comparedwith competing prices ranging from 4 to 6 all the way down to 1to 5.

Witness the following top prices about Best Actress: 11 to 10Julia Roberts, 12 to 1 Joan Allen, 20 to 1 Juliette Binoche, 20 to 1Ellen Burstyn, 20 to 1 Laura Linney. This makes up to a marginin favour of the bettor of 30.4%. The book is what is termedoverbroke. It is, in effect, the equivalent of a bookie offering 15 to8 against both horses in a 2-horse race.

A similar pattern emerges in an examination of the best openingprices about Best Actor: 7 to 4 Russell Crowe, 15 to 8 Tom Hanks,10 to 1 Geoffrey Rush, 20 to 1 Javier Bardem, 20 to 1 Ed Harris.This adds up to a margin in favour of the bettor of 11.2%, orabout the same as 5 to 4 against each of two horses.

For the record, Best Picture that year opened at 4 to 5 Gladiator,5 to 1 Traffic, 8 to 1 Crouching Tiger, Hidden Dragon, 20 to 1 ErinBrokovich, 40 to 1 Chocolat. That favours the bettor by a margin of9.5% (equal to about 6 to 5 against each of two).

Those thinking this was a one-off need look no further thanthe equivalent market for 2002. The favourite for Best Actress thistime was Sissy Spacek, at a best 4 to 5, and as short elsewhere as1 to 2. Meanwhile, Nicole Kidman was as long as 5 to 1 and asshort as 9 to 4, Judi Dench as long as 10 to 1 and as short as 11 to4, Halle Berry as long as 12 to 1 and as short as 3 to 1. Bringing upthe rear was Renee Zellwegger, as long as 25 to 1, and as short as12 to 1.

At best prices, the bettor has an in-built margin of about 6%.Even better news was available in the Best Supporting Actress

category, with the favourite Jennifer Connelly available at 6 to 4

123

BETTING ON THE OSCARS

(and 8 to 15), Marisa Tomei at 5 to 1 (and 9 to 2), Helen Mirren at6 to 1 (and 5 to 1), Maggie Smith at 6 to 1 (and 5 to 2), and KateWinslet at 12 to 1 (and 3 to 1). Once more the margin favouredthe bettor, this time by 7%.

If that was not enough to guarantee a small profit to no risk,you could add to your wad by turning to the Best Actor category,which yielded an advantage to the bettor of 10%, and to BestDirector for a massive 16%. Don’t believe that 16%? Well, here itis: Ron Howard at 11 to 4, Robert Altman at 7 to 2, Peter Jacksonat 4 to 1, Ridley Scott at 10 to 1, and David Lynch at 16 to 1.

The reality, of course, is that some of the better prices movedwithin minutes of the start of play. Even so, I can personally vouchfor the fact that they stayed around long enough to turntheoretical advantage into practical effect, albeit often tosomewhat limited stakes.

Still, there is something else we can derive from prices such asthese, which allows us an opportunity to profit more generally,and that is what they tell us about the true probabilities of eachoutcome.

The key question we need to resolve in these circumstances iswhere the best forecast lies. Is it with the odds-setter who takes aposition out of line with that quoted by others? Is it with themost conservative market-maker (the shortest odds quoted)? Isit, perhaps, with the average of all those who decide to venture aprice about the outcome?

There are different ways of approaching this question, but onemethod is to examine the size of the margin (or over-round)implied in the odds offered by each of the bookmakers. In acompetitive market, there is some economic justification forexpecting that the firm offering the most generous odds overall(the lowest over-round) is the firm that is most confident in itsestimates. Likewise, the firm with the highest over-round is theleast confident.

The bookmakers in my sample offering the best odds overall(i.e. the lowest over-round or margin) in 2001 for the categoriesof Best Picture, Best Actor and Best Actress were as follows:

124

BETTING TO WIN

Most generous odds for Best Picture: 117.97% (Paddy Power)

Most generous odds for Best Actor: 115.57% (Paddy Power)

Most generous odds for Best Actress: 110.07% (Coral)

The odds offered by each of these bookmakers respectively abouttheir favourite in each category were as follows:

Best Picture: Gladiator (4 to 5)

Best Actor: Russell Crowe (7 to 4)

Best Actress: Julia Roberts (11 to 10)

Applying this system to the 2002 nominations, again using asample of bookmakers, produced the following result:

Most generous odds for Best Picture: 117.2% (Bet Direct)

Most generous odds for Best Actor: 118.8% (Bet Direct)

Most generous odds for Best Actress: 110.5% (Sporting Odds)

The odds available with these bookmakers respectively were:

Best Actress: Sissy Spacek (4 to 5)

Best Actor: Russell Crowe (8 to 13)

Best Picture: A Beautiful Mind (4 to 5)

Although these were the favourites with each of the bookmakersoffering the best odds overall, in 2002 there were better odds availableelsewhere about the selections in the Best Actor and Best Picturecategories.

Thus, Russell Crowe was available at 5 to 4 with Victor Chandler,who were less generous about the competition. The best oddsavailable about A Beautiful Mind came courtesy of Ladbrokes, whooffered a standout 5 to 4, compensated by the shortest odds aboutLord of the Rings.

So there we have it. Using this system generated the followingbets:

125

BETTING ON THE OSCARS

2001Best Actress: Julia Roberts (11 to 10) – Result:WON

Best Actor: Russell Crowe (5 to 4) – Result: WON

Best Picture: Gladiator (4 to 5) – Result: WON

2002Best Actress: Sissy Spacek (4 to 5) – Result: LOST

Best Actor: Russell Crowe (5 to 4) – Result: LOST

Best Picture: A Beautiful Mind (5 to 4) – Result: WON

Worth trying next time?

126

BETTING TO WIN

Chapter 22: Second Impressions Are

Sometimes Better

First impressions, it is said, count for a lot in life. In the world ofhorses, odds and betting, they count for a whole lot more. It is thereason why so many promising two-year-olds retain ‘short-odds’status however poorly they subsequently perform on the big stage.It is what I sometimes call the ‘Tenby effect’, to describe the relativelysmall Caerleon colt whose mediocre performances as a three-year-old did nothing to dissuade punters from taking odds-on abouthim for the 1993 Derby. After trailing in tenth, all the usual excuseswere made, and it was really not until the end of his career thatpunters were finally persuaded that the odds about him had alwaysbeen woefully short.

Those with somewhat longer memories may think instead of thegreat hope of the Aga Khan for that wonderful 1936 season (no, Idon’t remember it personally), the much-hyped Bala Hissar.Beautifully bred, he was introduced from an early age to admirersas ‘the colt by Blandford out of Voleuse, the dam of Theft’. Blandfordwas the Sadler’s Wells of his day. Despite losing his first outing as atwo-year-old, Bala Hissar was soon installed as Derby favourite, andhis woeful performance in the 2000 Guineas did little to put off thefaithful. After all, he had the breeding, but when punters flocked tohis trainer, Frank Butters, to ask about his ‘Derby horse’, they weremet with the sharp rejoinder: ‘I should like you to tell me first whichis my Derby horse!’

In the event, Mr. Butters and the Aga Khan were to triumph afterall in the 1936 Derby, but instead with the grey who was apparentlynot bred to stay, Mahmoud. Unlike his stable-mate this was a horsewhich had shown form on the track, when losing the Guineas byjust a head despite being left at the start.

Form on the track has always been, it seems, under-rated in themarket compared to form at the stud farm.

It’s an opinion worth bearing in mind, perhaps, before followingthe hype with your hard-earned cash.

127

SECOND IMPRESSIONS ARE SOMETIMES BETTER

Chapter 23: The Problem With Miss

World

What were the true odds that robbers would attempt to steal MissWorld’s crown on the night of the 2000 annual pageant, and fail? 50to 1, according to one leading bookmaker offering odds in advance ofthe first pageant of the Millennium. Identical odds were on offer abouta fight between two of the contestants breaking out on TV. 50 to 1about the totally bizarre? If you think that’s a shade generous, thenso do the bookmakers, since 25 to 1 was the best you could get abouta contestant losing her bikini top. Still, to those looking for a quickturnover, there was still the 100 to 1 available about one of the loveliesrevealing on the Channel 5 show that she was, after all, a man.

A bit of fun, perhaps, and should be taken in that spirit, but somespecials really can offer value to the discerning bettor, as a numberof bookmakers have learned over the years to their cost.

The Miss World contest is potentially such an occasion, but thereason for considering it here is what it can tell us about a certaincategory of events more generally. To see what I mean, let us for thesake of argument divide betting propositions into three categories.

The first category is the sort of bet where the probabilities areclearly defined. The chance of ‘heads’ on the toss of a coin, or thered on a roulette wheel, or a ‘6’ on the roll of a die. The probabilityof each can be objectively measured (assuming the equipment isnot bent or faulty), and the bookmaker is able to shade the odds inthe knowledge that in the long run the betting public will lose.

The second category is the type of bet with a wealth of establishedform, like the performance of a horse in a handicap or a footballteam in the Premiership. Here the bookmaker has to make somejudgement of the probabilities, but is aided by techniques of analysisbuilt up over years of studying the impact of past form on futureoutcomes.

The Miss World contest is very much in the third category. The‘form’ of the contestants, as defined in the conventional betting senseat least, is strictly limited, and the consequence is that the odds onoffer from different market-makers can be very different. Beauty is

128

BETTING TO WIN

very much in the eye of the beholder, and in this case the bettor is ata clear advantage over the market-maker in being able to review theopinions of a diverse spectrum of those doing the beholding.

The conventional tools of betting analysis can and should still beapplied, however, and in particular the well-established ‘favourite –longshot bias’. This is the regularity, explained earlier, which indicatesthat, in the absence of other information, a blind strategy of bettingat shorter odds outperforms in the long run a strategy of betting atlonger odds.

Almost everywhere that it’s been tested this phenomenon applies,and there is no reason to expect that the Miss World spectacularshould be any exception.

The last time I studied the form of Miss World was at that 2000pageant, when Miss India was the general favourite at best odds of7 to 1. On this basis, she should have been my choice. The problemfor me was that another Miss India had won the year before, andsurely a representative of the same country wouldn’t win again (yes,I was temporarily affected by that dreaded affliction which I havediagnosed in all too many others – the ‘gambler’s fallacy’). And so,bowing to the fallacy, but cognisant of the bias against longshots, Iplumped for an alternative strategy based around the eliminationfrom my calculations of any entrant rated by a mainstream market-maker as longer than 20 to 1.

On the basis of a wide selection of all odds on offer, this left at thetime of selection only the following: Miss Colombia, Miss Costa Rica,Miss South Africa, Miss USA, Miss Venezuela, and Miss India herself.

To distinguish between the relative merits of these contenders Idecided, in the absence of superior information, to calculate theaverage probabilities as implied in the odds made available by thedifferent bookmakers.

This gives the following guide to the approximate average odds(best odds in brackets):

Miss India 6 to 1 (7 to 1); Miss Venezuela 6 to 1 (7 to 1); Miss SouthAfrica 13 to 1 (16 to 1); Miss USA 14 to 1 (16 to 1); Miss Costa Rica 15to 1 (20 to 1); Miss Colombia 18 to 1 (20 to 1).

The solution followed directly. Take the relatively ‘generous’ 20to 1 (average price 15 to 1) available about Miss Costa Rica, 22-year-old Cristine de Mezerville Ferreto. She might not have the best

129

THE PROBLEM WITH MISS WORLD

chance, but there was good reason for claiming value.The alternative strategy was instead to bow to the dictates of the

favourite – longshot bias. In other words, back Miss India.And the winner? Well, Miss India, of course! And the losers? Miss

Costa Rica, naturally, and me.

130

BETTING TO WIN

Chapter 24: Looking For Patterns

Did Gordon Banks’ stomach upset inadvertently cause the defeatof Prime Minister Harold Wilson’s Labour Government in 1970? Mr.Wilson certainly thought so, for who can believe that West Germanycould have come back from 2 – 0 down with minutes left to ejectEngland from the 1970 World Cup, if Banks had been in goal. Theman whose form was so inspired that he stopped that Pele headerwould surely have proved more than a match for the predatoryGerd Mueller. With the Germans out of the way, England wouldhave gone on to exact revenge on the Brazilians and successfullydefend the Jules Rimet trophy. A nation in jubilant mood wouldthen have re-elected Mr. Wilson with a thumping majority. At leastthat’s how one man saw it.

Establishing a link between events which have no obviousrelationship has long fascinated astrologers, of course, but in recentyears the advent of computing power has brought such analysisever more into the realm of conventional forecasting.

A statistic identified by the ABC sports team, just prior to the 2000US Presidential election, is just one example. The statistical wizardsemployed at ABC noticed an amazing relationship between theperformance of the Washington Redskins in their last home gamebefore a US Presidential election and the outcome of the election. Ifthe Redskins win, the candidate of the party already in office (inthis case, Al Gore) wins. If they lose, the candidate of the mainopposition party wins (i.e. George W. Bush). A track record whichhad proved itself fifteen times out of fifteen. The game was mightyclose, and the Redskins lost narrowly, some would say as a result ofa dodgy decision. The rest you know.

The problem with this sort of statistic is that it smacks a little tooclosely of the famous ‘Superbowl Effect’, identified and published byProfessors Krueger and Kennedy in 1990 in the Journal of Finance. TheSuperbowl Effect was the link between a team from the National FootballConference (NFC) winning the Superbowl and an improvement inthe following year ’s stock market. Unfortunately for Krueger andKennedy, after 23 consecutive forecasting successes between 1967 and1989 it went down in 1990, along with the Denver Broncos.

131

LOOKING FOR PATTERNS

The reason for these freak relationships is linked to the issue ofso-called data-mining, i.e. the idea that data is mined by researchersuntil an apparent trend is found. The words of Nobel Prize-winnerFischer Black, when such a freak statistic was brought to hisattention, are instructive.

‘It sounds like people searched over thousands of rules till theyfound one that worked in the past. Then they reported it, as if pastperformance were indicative of future performance. As we mightexpect, in real life the rule did not work any more.’

Modern computing power makes it easier than ever to come upwith such patterns. They are what statisticians term spuriouscorrelations.

There are, of course, plenty of real patterns which can helpimprove our forecasts, and it is important to distinguish the realthing from the impostor.

The first thing is to ask if it makes any sense at all. If it does, whatforecasters do is to test the pattern in a different sample to that whereit was found. If it still exists, test it again and again and again untilyou are sure beyond any reasonable doubt. If it makes no sense,you have to test even harder, and try to come up with an explanation.Proving the incredible is just a little bit harder than proving themerely unlikely.

132

BETTING TO WIN

Chapter 25: The Motivation Thing

Ever since David slew Goliath the shock upset has become a well-documented part of the world’s unfolding history. Less clear is whyit happens, and how we can be among the élite band of those whocan predict when it will.

Some look to emotions and motivation for the answer. Revenge,after all, is a powerful emotion; survival is stronger still. It was aweek in May, 2000, that got me to asking the question.

It was the week that European Champions Manchester Unitedfell at Old Trafford at the hands of the team they had vanquished towin the treble, Bayern Munich. It was also the week that languishingDerby did the same to them in the Premiership.

Revenge, emotion? How else can we explain why a Man Utd sidewhich had lost just one home Premiership game all season, scoring49 times and conceding just 11 goals, could lose to an injury-weakened Derby outfit by the only goal of the game. It wasn’t theirbest side, admittedly, but the team starting with Beckham, Cole andSheringham would at any other time surely have put at least a couplepast the ailing Rams. It was even more remarkable in a May weekend,which like so many May weekends before, had witnessed a feast ofgoals across the breadth and length of the Premiership.

The theory of motivation does sound plausible, but could thatexplain why on the same day a Coventry side much more desperatefor points than Derby could squander a two-goal lead to lose to AstonVilla. Could it explain Middlesbrough’s indifferent display away toan already relegated Bradford? Could it explain why a points-famishedManchester City could let slip two late goals to Ipswich?

In reality, a dispassionate analysis of the cold statistics of thesethings reveals no apparent correlation between the need to win andthe actual win. Indeed, the evidence suggests that teams that needthe points desperately do no better on average than teams that needthem somewhat less desperately, or even than teams that don’t reallyneed them at all.

Instead, better teams usually tend to outperform lesser teams,after allowing for home advantage, and this is unaltered by therelative need of these teams to win.

133

THE MOTIVATION THING

This is not to say that in any individual circumstance those witha special insight or talent are unable to spot likely upsets. This isperfectly possible, and is quite consistent with the cold statisticalapproach.

The reason that both approaches can work is that what is true onaverage is not always true in the particular. To clarify, a team thatneeds the points playing a team that doesn’t is in general likely tofare no better than if the roles were reversed. In any particular case,however, the exception may prove stronger than the rule. Findingthat exception requires rare insight, or luck.

Without this insight the better play, according to the statistics, iseither to ignore the ‘need to win’ factor, or if anything to go theother way, on the basis that the odds will on average over-reflectthe common misperception.

Outside the world of football this phenomenon has been treatedwith less statistical rigour. Instead, we are treated to an array ofanecdotal evidence. Tales abound, for example, of the ‘hungry’ boxer,with limited skills, who overcomes the mighty champion. BusterDouglas knocking out Mike Tyson, Hasim Rahman despatchingLennox Lewis, are two obvious examples. In truth, however, thesereally are exceptions, and the vast majority of outcomes reflect therelative class of the opponents.

Even so, intuition would indicate that motivation might play amore influential role in a one-off fight between two individuals thanin a game between two teams who are playing every few days. Thereason may lie in the fitness aspect so vital to modern sport. Amotivated boxer is likely to be fitter and sharper, and this can bedecisive. This is surely less critical when considering the relativemerits of two teams of professional football players at the top end ofthe game. Even if one side is fitter than the other, this is likely to bepublic knowledge, and well and truly factored into the odds.

Similarly, in an individual sport like boxing or tennis, an off-daycan spell disaster. The famous first round Wimbledon exit of MartinaHingis to unseeded Jelena Dokic is a well-known example. Not untila high profile Florida court case hit the headlines did we hear tell ofthe effect that a stalker might have had on the outcome!

And so back to those twin motivations of revenge and survival,and how they can help us in unravelling the value in that Champions

134

BETTING TO WIN

League Final or that opening Test Match. Who will be most up forit, who will be seeking to even the score? Will it be this; will it bethat?

In general, you needn’t worry too much about any of thoseconsiderations. Instead, just follow the form. And if you get it wrong,don’t get emotional and don’t seek revenge. Keep cool, keep calmand get even.

135

THE MOTIVATION THING

Chapter 26: Noise

More than 30 years have elapsed since a computer program workedout the result of a mythical encounter between Rocky Marciano andMuhammad Ali, but the 2001 ‘virtual’ Champion Hurdle, ChampionChase and Gold Cup stemmed from the same idea. If it isn’t goingto happen in real life (and with the abandonment of that year ’sCheltenham Festival it wasn’t), we sure want to know what wouldhave happened if it had. A sort of sophisticated Pools Panel, youmight say, but a whole lot more fun. For the record, Rocky knockedout ‘The Greatest’ in the 13th round of a bruising encounter, whileIstabraq, Tiutchev and Legal Right showed what happens whenhard logic is showered with a sprinkling of stardust.

All good fun, but behind the idea lies something which may beworth considering in implementing any betting strategy. It boils downto what economists call ‘noise’, or the tendency of any closed systemto be afflicted by random factors which cannot be predicted in anysystematic fashion by a model. Over the long run, noise tends toeven out, so that a good model will overestimate and underestimateperformance indicators in equal measure. The predictions will beunbiased. In the short run, however, the more noise in the model,the more risky any bet based on the prediction will be.

Let me give an example. If there is no noise at all, i.e. no randomvariations, then any mistakes in predicting the outcome of an eventare the sole fault of the forecasting system. An example is the effectof gravity on an apple. A model of the way the world behaves willpredict that if you drop the apple it will fall. There is no noise in theprediction, apart from the minor effects of air resistance and thelike. However. now take a race like the Grand National, and theinfluence of random factors reduces the confidence in any estimate,however sophisticated the forecasting system. It may get it right inthe long run, but the long run can be a very long time, and certainlya lot longer than the time it takes to run the race.

Followers of forecasting systems must, therefore, take account ofthe amount of noise in the system before deciding how much tostake, and how much confidence to invest in any individualprediction.

136

BETTING TO WIN

In horse racing, chases are likely on the whole to be more noisythan hurdle races, which are in turn likely to contain more noisethan five-furlong sprints. Of course, the form may be moreestablished in the chase than in the novice sprint, but we are talkinghere of random influences additional to that contained in the formbook, such as it exists. Again, some sports are likely to contain morenoise than others. A game where points are scored quite regularlymight be expected to produce a less random result in a given periodof play than a game where points (or goals) are scored infrequently.In this sense, basketball, for example, is likely to be less noisy thanfootball. Generally, noise is likely to contribute a greater effect to theoutcome where the frequency of scoring is low, and when the timeperiod over which the game is played is short.

Sometimes, of course, there is no straightforward way of thinkingabout the effect of these random influences. Take the Boat Race asan example. Those who can remember or know of the sinkings, themost recent of these (at least during the race) being the demise ofthe Cambridge boat in 1978, will guess that there is quite a lot ofnoise in this event, in every sense. In fact, the favourites tend to winquite regularly, which suggests that apart from the odd moment ofmayhem the smooth progression of each team to the winning posttakes place in a reasonably orderly and predictable fashion. Thechoice of station, the undulation of the waves, the baying studentson the banks, really doesn’t seem to cause as much noise as evenone small cox.

Which means that you may be on to a good thing by siding withthe favourite, to win by the margin implied in the odds. Thus, 4 to 6Oxford should mean an Oxford victory, just. In which case youshould back them to do so – just. Try it some time.

137

DRAW BIASES

Chapter 27: Betting On A Leadership

Election

‘A week is a long time in politics,’ once declared Harold Wilson. Itwas a sentiment echoed in the triumph of Kenneth Clarke in thefinal ballot of Conservative Members of Parliament, prior to the 2001ballot of all party members nationwide.

That victory marked the latest episode in a turbulent few weekswhich started out with the Member for Rushcliffe being quoted atapproaching double figures in the days before the announcementof his decision to stand. That decision, announced on Tuesday, 26June, came to the surprise of many. Victor Chandler, for example,chopped him from 8 to 1 to 100 to 30 after the announcement. Thesurprise was in part the result of the low profile of the man whoaccording to press reports had been leading an ostensibly Trappistexistence over the preceding few days. No surprise to yours truly,however, who as a fellow member of Notts C.C.C., was engaginghim in an illuminating chat after stumps on the preceding Saturday.

The strangest thing about the leadership election campaign priorto the ballot of MPs, however, had not been the swings androundabouts of fortune for the contenders, but rather therollercoaster of inconsistent odds offered by the market-makers.Odds about Clarke shortened markedly on the morning of the finalballot, for example, until only Paddy Power were willing to go anylonger than 2 to 1, and then only to a maximum of £14. By the timeof the bell, Blue Square were offering 4 to 7 Duncan Smith, 3 to 1Clarke, and 7 to 2 Portillo. Given Portillo’s so recent quote of 1 to 3,this was yet another example of the cricket paradox, whereby it isall so often possible to bet on every eventuality at a good shade ofodds against if you just have the patience to wait.

In the end, then, all we learned about the skills of political market-makers from this election is that they cottoned on to the fact thatMr. Portillo was losing ground.

No sooner was the result declared, however, than we were backto that same rollercoaster, this time equipped with a strange flip-back. Just look at the logic. If Iain Duncan Smith was such a hot

138

BETTING TO WIN

shade of odds-on to win the election outright before the poll of MPs,how come that within moments of the declaration, all firms (barCoral – who adjusted fast) were now quoting Clarke as the hotfavourite. After all, the odds shortly before the declaration, in thefixed-odds and spread markets, were generally of a mind thatDuncan Smith and Clarke would make it through, and yet Mr DS atthat time was still being quoted at odds as short as 1 to 2.

Perhaps the logic lies in the belief by some that Duncan Smithwould easily beat Portillo, but not Clarke, and that the chance ofPortillo coming through to the final play-off skewed matters.

Perhaps the fact that Clarke scored more votes than expectedamong MPs had an impact. The problem with that analysis is thatConservative MPs make up only one in 2000 of the partymembership. That apart, it was always expected that all threecandidates would poll somewhere between 50 and 60 votes. So thedifference between the actual and expected vote was marginal atbest.

All we can say with confidence is that the odds-setters themselveshad very little confidence in the odds they set about internalConservative Party matters. Indeed, this inability of pundits andmarket-makers to fathom the mind of the Conservative Party hasechoes dating back to the accession of Alec Douglas-Home to thejob in the early 1960s. They were no better informed in 1997, whenthe Clarke-Redwood pact led to a drastic shortening of the formerChancellor ’s odds, just at the time as the Tory MPs were recoilingso heavily from the deal as to give the prize to the Member forRichmond.

I decided, therefore, that this was one event where following theodds was for mugs. Instead I used my so-called ‘common sense’ inguessing that a bigger name would always appeal more to the rankand file. In other words, I was banking on a glorified version ofwhat American political pundits tend to call the ‘name recognition’factor.

The rest, as they say, is history!

139

DRAW BIASES

Chapter 28: You Can’t Beat

Reliability

Who would you back over a mile and four furlongs – Lester Piggotton a carthorse, or John McCririck on Nijinsky? The sad thing forthose of a Corinthian spirit is that in real life it is likely that Big Macwould be astride old Steptoe’s nag.

The best jockeys attract the best horses, the best football clubsattract the best players, and the best drivers attract the best cars.Which is why there are rarely more than two or three real contendersfor the Formula 1 World Drivers’ Championship.

Now assume for the sake of argument that there is little to chooseon the tyre front. Little either to choose in latent ability. If thesefactors were even, that doesn’t mean that equivalent odds representequivalent value. For that we need to add reliability and predictabilityto the balance, for therein lies the key to value.

What Formula 1 fans call reliability, economists and statisticianscall volatility, or risk. Moreover, the very same economists haveconstructed a model which estimates what extra return is requiredto compensate for an additional dose of risk. This is sometimesknown as the risk premium, and makes up part of what is called the‘Capital Asset Pricing Model’.

Applying the essentials of the model to Formula 1 racing, thismeans that a conventional investment strategy would require betterodds to side with an unreliable driver or car, even if there were nodifference in how well they might be expected to perform on theaverage. This is why we need to examine both the mean (average)performance and the variance (the dispersion of results aroundthis average). We want to maximise the former, but minimise thelatter.

In other words, if the McLaren and Ferrari cars were equally good,and would both be expected on best estimates to win six races, weshould require better odds to side with the car that achieves this inthe less consistent way. The same goes for the driver.

All too often, of course, the faster car is also the most reliable, thefaster driver the same. In these cases we can feel all the more

140

BETTING TO WIN

confident about taking short odds. So it’s Schumacher to win in theFerrari. That is what I call the appliance of betting science. Othersmight, of course, call it ‘déjà vu’.

141

DRAW BIASES

Chapter 29: Cricket Betting

Backing the outsider of three at the racetrack never was the wisestadage in the pantheon of betting folklore, but there are times whenbacking the ‘insider ’, or shortest priced of three, is sometimespositively worse. I refer to those occasions when the bookmakersconspire to offer the draw in a cricket match as the hot favourite.The first time I was asked to look at this was just before the SecondTest at Old Trafford in 2001, between England and Pakistan. Onecould accept that the weather forecast was not glorious, and thatthe expectation of a low, slow surface at Old Trafford was expectedto favour the bat. Still, an offer of evens about the draw, by SurreySports, was scandalously short, and even the standout 13 to 8 onoffer with the Tote was notable more for its relative, rather than itsabsolute, generosity.

The fact is that the odds about the draw in a cricket match are, onaverage, the worst possible of the three options, and yet howevershort the bookies go the evidence suggests that the madding throngof punters will follow them in. It is as if punters believe that cricketmarket-makers have some special insight into the weather sharedby none since the pre-hurricane days of the BBC’s Michael Fish.Indeed, a cramped shade of odds about the draw is taken toforeshadow the near certainty of five days of torrential rain orponderous black clouds.

Life isn’t like that, but it doesn’t take much to encourage thebookies in their ways. Despite a respectable run rate on the firstday, bolstered by 53 boundaries, the chances of England were outto 7 to 1, as the draw shortened to 1 to 2. And still the money wentdown. By the third day, it was the turn of Pakistan to lengthen to 7to 1 and the draw stood at 1 to 3. Sunday saw the prices levellingoff, at 5 to 2 England, 6 to 4 Pakistan and 13 to 8 the draw. And justfor the record, the final day opened with England still at 5 to 2 topull off a victory rated no more than a one in eight chance after thefirst day’s play. Pakistan were on offer at 7 to 2 and the draw a bareshade of odds on, at a best-priced 10 to 11.

Swings and roundabouts, as they say, and of course you couldhave bet on all three outcomes for a very healthy profit at some

142

BETTING TO WIN

point. You could indeed, but hindsight is a wonderful thing, andyou cannot be sure to repeat the trick on any given future occasion.Even so, there is some sense in the argument, inasmuch as cricketodds do tend to swing a little too sharply in response to the latestday’s play, at least if you play at the best available odds across thespectrum of bookmakers.

Which brings us to a couple of rules of thumb which might proveuseful when trading on the cricket.

First, ignore the draw. However good the odds seem, it is veryunlikely that they represent value. Too many people have plungedtoo much, too often, on the draw, for the bookies not to take noticeand advantage accordingly.

Second, the best value is likely to obtain during the course of thematch, as one or more bookmakers offer a standout price which canonly represent an over-reaction to a single day’s action.

Sometimes, however, your heart really must rule your head. Onthe assumption that most readers follow the home side, why nothave a fiver on England at the start of every Test to win the series?

Conscience clear, you can now wait for day two. And watchcarefully, for there is one thing for sure in a Test series: whoeverwins at the cricket, there should be only one winner in the battlebetween sophisticated bettor and bookmaker. And that winnershould be you!

Part Three: Beating The Tote

Chapter 1: When Betting Longshots,

Getting What You’re Given Really

Can Be The Best Option

It’s a balmy summer evening at Newmarket’s July course, as youstudy the horses parading before the next race. That mount ofDettori’s looks particularly well. So you walk the few yards to therow of betting booths representing the Horserace Totalisator Board,and stake your last fiver that the favourite will indeed frank theform. Or do you? Well, it’s definitely a lot more convenient thantrekking hastily back to the bookies’ boards lining the long straight,and you’re certain to get on in time. You can’t be sure of the priceyou’re getting, of course, but there’s no reason to expect a bad deal,is there? According to standard economic theory, you would be right.The so-called ‘Efficient Markets Hypothesis’, to be precise, suggeststhat the returns to investments of similar riskiness placed in any ofa number of complementary markets should even out over time.Whether you invest with the bookies or the Tote, therefore, youmight expect the same return. Otherwise clued-up investors shouldstep in and mop up any value until it disappears. But do they? Inmost markets, the weight of evidence suggests that they do. Not so,it seems, at the racetrack. There are systematic differences in whatyou can expect to earn for every pound staked, depending onwhether it ends up in the Tote till or the bookie’s satchel, and thisdifference is especially pronounced at long odds. This so-called‘anomaly’, as finance specialists are apt to call it, has been guessedat for years. The first large-scale academic study to prove what wasalways expected, however, can be traced to the work of twoAmerican academics, at the time operating out of Universities inChicago and Kentucky.

Paul Gabriel and James Marsden examined 1427 flat races fromthe 1978 British flat racing season. Using a range of statistical tests,they confirmed that the Tote returned on average a higher amountof money to a unit stake than the bookmakers, and that this wasespecially so early in the season.

146

BETTING TO WIN

Since then a number of other studies have been published, eachsupporting the finding of a higher average return at Tote odds thanat starting price. These later studies do, however, suggest that thedisparity is concentrated at odds of 5 to 1 or so and above, and thelonger the odds the greater the advantage to the Tote. Below 5 to 1,the disparity is less clear. There is also the added advantage withthe bookies that you might just beat the returned price if you shoparound.

So why bother with the bookmakers when betting the longshots?The answer lies in the nature of the beast. That 10 to 1 shot mightend up favourite, but at the moment it’s good value and the bookieswill allow you to lock in your advantage. Not so the ‘nanny goat’(Tote). If you realise that the 10 to 1 is good value, no doubt there’llbe plenty of others who realise the same thing. By the time the goodthing is not such a good price, it’s too late. You’ve paid a fixed stake,but you’ll get back what you’re given.

The lessons to be learned from all this research are clear enough.Unless you know the price currently available is that little bit toobig, steer your cash towards the Tote if you expect the starting priceto be in the region of 5 to 1 or longer. Below 5 to 1, and you’respoiled for choice, although there is some evidence that money beton sure things, especially when they are carrying popular jockeys,is best placed in the bookies’ satchels. Oh yes, and if you do spotvalue, especially early on, seek it out and lock it in. If you’re right, itwon’t last long, no matter how you bet.

147

DRAW BIASES

Chapter 2: Steaming Into The Tote

It’s Champions day at Newmarket, and with one race left on thecard you are left pondering your options. Suddenly you are thrownback in time to ‘Arc’ day, at Longchamp – Sunday, 1 October 2000.

In particular, your mind is drawn to the fourth race of that day,the Prix de l’Abbaye de Longchamp – Majestic Barrière. Your initialpick, Bertolini, stood clear above the others as a general favourite atmorning prices, followed (at best prices) by Pipalong, SampowerStar, Superstar Leo and Primo Valentino.

After these came a game Irish sprinter named Namid, trained byJohn Oxx and ridden by Johnny Murtagh. Indeed, Chandler, Coraland William Hill were of like mind in offering 10 to 1 to any takers.In the event there were so many takers that you were unable to findanything better than 100 to 30 by the time that you were ready tofollow the money with your own hard cash. Unless, that is, youturned to the French equivalent of the Tote, the so-called parimutuel.Always available at the equivalent of about 6 to 1, the steamer in thebetting offices was, at the off, still available on the parimutuel at almosttwice the price you could get with the bookies. In the event, it wasreturned at 5.2 to 1 (1.8 to 1 a place), after storming home to a facilevictory over Superstar Leo and Pipalong. For the record, the early-morning favourite, Bertolini, was beaten home by six others.

What makes this example so interesting is that the parimutuel pricewas rather lower than you could have obtained with bookmakersearlier in the day, but significantly better than the final price available.

Studies of the phenomenon of steamers have been published insome of the foremost academic journals in the world, and they allreveal the same thing, explored earlier in this book. A horse whichshortens in the market is a better bet, at any given price, than onewhich does not. For example, a horse which shortens from 10 to 1to 100 to 30 (like Namid) represents better value, even at the 100 to30, than an equivalent animal which had always been available at100 to 30. That’s the good news.

The bad news is that the final price is, on average, not good enoughto translate this observation into a long-term profitable bettingstrategy. By backing the steamer, all you can expect is to lose

148

BETTING TO WIN

significantly less than by backing a horse that had not shortened inthe same way. Even so, at the original price, or at any price muchabove the final price, you would indeed expect to turn loss intoprofit in the long run. That’s a lesson which followers of the Namidexperience should learn. At the 100 to 30, the value had gone. At the5.2 to 1 the value was, in principle, still available. By using theinformation contained in the movement of the bookies’ odds, andthen searching out the French Tote odds, you were well ahead invalue terms before the first sprinter exited the stalls.

Wake up now from your fleeting journey through time, and youare back at Newmarket considering which of the 30 runners liningup for the seven-furlong NGK Spark Plugs Handicap is likely toromp home first. It’s a difficult choice until you remember the simplelesson you learned at Longchamp. You are looking not for the mostlikely winner but for the best value. After all, while value doesn’talways translate into winnings, over the long term it surely will.

This simplifies the choice considerably, and brings straight intothe frame the Richard Hannon-trained three-year-old, SocialContract. Suited on form to the distance, well drawn, and wearingblinkers for the first time, the 25 to 1 offered generally with thebookmakers was if anything slightly generous. Yet this same horsewas readily available at more than three times that price on the Tote,and there were few signs of any convergence with the board prices.

In the event, the blinkers worked, and five-pound claimer, PaulFitzsimmons, eased home to land the odds, giving backers on theTote a final payout of 79.4 to 1.

149

DRAW BIASES

Chapter 3: Betting On The US Tote

My own on-track experience of US horse racing started at ChurchillDowns, the home of the Kentucky Derby. It costs only $2 (little morethan a pound) to gain admission. It’s exciting, colourful, clean andcomfortable. Best of all, there is a simple well-established way toearn a modest and consistent profit. I’ll tell you about it.

First, the basics: in the USA, there are no bookmakers at theracetrack, only the parimutuel (what we call the Tote). You can betto win, to place (must come in the first two), or to show (must come inthe first three). There are also all sorts of so-called exotic bets, likethe ‘trifecta wheel’, where you bet on six horses to finish in the firstthree in some predetermined order. You can even bet ‘odd or even’on the card number of the horse you think will win. It’s like a livingroulette game, and the take, at only 5%, is little more.

There is also a way in which it may be possible to use the odds inthe win pool to take a profit from the place and show pools.

The method relies on a mathematical formula (originally calledthe Harville formula), which sounds complicated but is quite simpleto operate. Essentially, it works like this.

First, find the win odds. These shift about, but because of the bigpools they don’t shift that much, and can be observed from thehuge electronic boards easily visible from everywhere at the track.Once you’ve identified these, the formula tells you what the oddsshould be that a horse will finish second or third. If the odds arebetter than they should be, you simply bet to place (to finish in thefirst two), or to show (to finish in the first three).

You can buy special calculators to do the sums for you, the mostfamous being William Ziemba (Dr. Z)’s special ‘Beat the Racetrack’calculator. Even without the calculator, there are some simple ruleswhich will help you on your way.

First, don’t back longshots, but instead concentrate on shorter-priced horses, especially if they are maidens (have not won before)or have not run for a while. Most punters at the US racetrack thinksuch horses will either win or come nowhere. This is the so-called‘Silky Sullivan effect’, named after the horse of that name whosecareer consisted of little other than stunning victories or ignominious

150

BETTING TO WIN

defeats. Not for Silky second or third place. If he couldn’t win, hewouldn’t try at all.

The more astute punters know from the statistics, however, thatSilky Sullivans are more rare than imagined or implied in the placeand show odds, and take advantage of this mispricing to reap smallbut consistent rewards.

So you have two choices. You could buy a racing calculator, andbet when the odds to place or show are too high. Otherwise, youmight look for horses priced at, say, 4 to 1 or less. If these horseshave not run for a while, or are maidens, bet on them to place orshow. If and when they finish in the first two or first three you will,according to the system, likely be returned a small but quiteconsistent fistful of dollars.

So keep your head, don’t risk too much, and you can have awonderful day out at the US racetrack – and leave with a walletwhich is bulging that little bit more satisfyingly than when youarrived.

151

DRAW BIASES

Chapter 4: US Horseracing On The TV

The wonderful thing about betting on American horse racing is thatyou don’t need to know anything about the horses to do creditablywell. Instead, much of what you need to know is contained in theabundance of odds displays everywhere in view at the racetrack.

There may be a limited number of readers of this book who havethe money, time and opportunity to turn up in person at GulfstreamPark or Golden Gate Fields or any of the host of other exoticallynamed racetrack venues. But you don’t need to. All this and more isavailable at the flick of the remote control. I am referring, of course,to the evening (UK time) racing from across the States, beamed liveand free each weekday to our shores.

Let me take you on a trip back in time, to Wednesday, 10 January2001, at 8.50 pm, to illustrate the message. Viewers of Sky Sports 2had just been treated to all that is best, and worst, about football, inthis case courtesy of Liverpool FC. Some wonderful touch play,squandered by abysmal finishing. Crystal Palace were still half agame away from a 2 – 1 victory. Those without the Racing Channelwere welcome to the half-time comments of Kenny Dalglish andco., while those blessed with the choice were able to turn straight tothe action from Race 8 at Laurel Park.

The ‘morning line’ estimates (similar to the trade press startingprice forecasts in the UK) about each of the entrants were clearly ondisplay, as well as the current odds. These are the so-called‘parimutuel’, or tote odds. In the UK you can bet on the Americanracing to win, or each way, or you can go for the Exacta (the firsttwo in the correct order). Other bets available at the racetrack arethe place bet (to finish in the first two), the show bet (to finish in thefirst three), the Trifecta (the first three in the correct order) theSuperfecta (the first four), and a host of other fascinating options,touched upon earlier.

Now let’s focus on the action. Holly Jolly, the no. 1 horse, waspredicted to start on the morning line at 2 to 1, but the weight ofmoney in the pool had shortened it up to evens by the time I decidedto miss the Sky panel’s expert analysis of the big match. By the off itwas 4 to 5, showing evidence of late money. The only other horses

152

BETTING TO WIN

to shorten from the morning line were no. 6, the bizarrely namedHunka Hunka Lori Z (from 7 to 2 to 3 to 1), and no. 4, Color MeHappy (from 3 to 1 to 5 to 2). Neither of these showed any signs ofthe late money effect. The odds about all the other horses hadlengthened since the morning forecast.

Turn now to the findings of Peter Asch and Richard Quandt, ofRutgers and Princeton Universities respectively. Employing datafrom 729 races at the Atlantic City racetrack, Asch and Quandtreported that the final parimutuel odds tended to be lower than thepredicted (or morning line) odds for winners. Moreover, in the caseof winners, the later in the betting period the money was laid, thestronger was this effect. Asch and Quandt proposed an explanationof this finding as the withholding of smart money (money bet bypeople with superior information) until late in the betting period,in order to avoid giving out market signals which could depress thefinal payout odds. As in any tote system, it is not possible to take aprice.

In other words, the best horses to bet on are those which haveshortened since being published in the morning line, and particularlythose which have shortened, or shortened further, late on. Add into this the well-established favourite – longshot bias, which appliesin the USA as well as the UK, and which dictates that the expectedreturn to bets at shorter odds tends to outweigh those at longerodds.

Applying all this theory and evidence at Laurel Park, the clearvalue lay in Holly Jolly, the 4 to 5 favourite, with some signs ofsupport for numbers 4 and 6. Of the latter entrants, there was littleto distinguish between them, each of which had shortened a littlein the market with no late surge of support.

These figures simply drive the bet: two Trifectas, one on horses 1,4 and 6 to finish in that order past the post. The other bet, of course,is the Trifecta on 1, 6 and 4.

In the event, it was no. 1, Holly Jolly who took the honours,followed home by nos. 6 and 4. The Trifecta payout was $31 to a $2stake, or in this case $31 dollars for minimal risk. Of course, this wastoo good to be true, in the UK at least, with the bookies unwilling toaccommodate further than the simple Exacta. So it’s two Exactas,coupling 1 with 6 and 1 with 4. Ah well, at least it’s a profit.

153

US HORSERACING ON THE TV

Now back to that Worthington Cup semi-final. How are thebookings going?

Part Four: Beating The Spreads

Chapter 1 : Profit Margins And

Profitability In The Spread Markets

The bookmaker offers you evens that the coin will land on its head,and evens that it will land on its tail. For a fair coin, those are fairodds, and in the long run both you and the bookmaker will comeout level. Of course, the bookmaker has to earn an honest crust,and so a notional profit margin is built into the odds, known as an‘over-round.’ We’ve illustrated how this works elsewhere in thisvolume, but it bears repeating. Imagine the bookmaker shades theodds about the head and the tail to 4 to 5 each. At these odds, youhave to bet £10 (£5 on heads and £5 on tails) to ensure a return of £9whichever comes up, i.e. £4 plus your £5 stake returned on thewinning bet. The over-round can be calculated as 10 divided by 9,i.e. 1.111 (or, more usually, 111.1%). You can apply the same logic toany number of options. The smaller the over-round the less in-builtadvantage the bookmaker has in the odds, and if the total is lessthan 100% the tables are turned. For fixed odds, this is a relativelystraightforward calculation. What about spread betting markets?

Take the example of a Test Match between England and Australia,and a performance index for the teams. Assume for a moment thatthe game has to end with one side winning, so that in the event ofa draw there would be a sudden-death shootout of some kind. Nowaward 25 points for the win, and none for the defeat. What wouldthe market-makers offer about each team? The favourite would beexpected to earn more than 12.5 points, and the stronger the degreeof favouritism the higher the expected points make-up. Say the quoteis 16 – 18 about Australia and 6 – 8 about England. The spread ineach case is two points, out of 25 points available. The over-round isquite simple to calculate in this sort of case. Whether you buy or sellEngland or Australia, it is given by calculating the share of the 2point spread in the total 25 on offer. This is 2/25, or 8%.

In fact, of course, the draw is a very real possibility, and the market-makers usually quote a spread based on 25 for a win, 10 for a draw,and zero for the loss. The problem is that because we don’t knowthe likelihood of a draw in advance, it makes the calculation of the

158

BETTING TO WIN

over-round much more complex.Assume now, for example, that IG’s opening quote about the

match was 16 – 17.5 for Australia and 6 – 7.5 for England. Say thatSporting Index, on the other hand, offered 15 – 16.5 and 6.5 – 8respectively. What we know, for sure, is that in these circumstancesthe over-round facing the bettor is much lower than the over-roundcontained in either of the individual sets of odds. This is becausethe bettor faces an effective spread of only 0.5 between the bestprice at which it is possible to buy Australia (16.5) and the best priceat which it is possible to sell (16), simply by shopping around.

Better still, whenever two quotes touch, the over-round facing thebettor is zero. For example, if IG offered a spread of 6 – 7.5 about England,and Sporting offered a quote of 7.5 – 9, the offers would touch at 7.5. Insuch a case, the effective over-round facing the bettor with access toboth offers is zero. The over-round implied in the odds offered by anindividual firm might be of academic interest, but no more than that.

In fact, the implicit over-round in the quote offered by any individualmarket-maker tends to be lower when the market is less volatile (ormore predictable), and lower again when there are fewer options. Thefinal of a match between two teams, one of whom must win, like theSuperbowl, on a 100 – 75 index, should as such represent particularlygood value. The worst over-rounds are at the start of such events as asnooker competition, where there are a large number of players. Herethe implied over-round reaches as high as 40%, and more.

Unless you’re confident enough to believe that you can beat thebookie and the over-round, the real trick is to choose those marketswhere more than one firm are offering quotes, and concentrate onthose where the spreads touch, or where the difference is relativelysmall. The bookings market (based on the number of yellow and redcards issued in a football match – 10 for a yellow, 25 for a red) is ideal, asit is quite common here to be offered, say, 44 – 48 by one firm, and 48 –52 by another. By having access to both these spreads (they intersect inthis case at 48), the effective over-round facing you is zero (and it’sdeduction-free). If you lose now, it’s because of bad judgement, or badluck, and not because of some inflated over-round which can be blamedon the bookmaker. In other words, it’s a level playing field, and yourfortune lies very much in your own hands.

159

DRAW BIASES

Chapter 2: Avoid The Christmas

Syndrome

Take two teams, packed with defensive talent, but imbued withrather less flair up front. Put these teams together and what can weexpect? The irresistible force meeting the immoveable object? Notreally. Rather, it’s likely to be a case of the immoveable object meetingitself. Maybe not the most entertaining spectacle, but certainly ahappy betting medium. Let me explain.

The value in such a match should lie in an awareness of thepropensity of bettors to succumb to the Christmas syndrome, i.e.the tendency to buy too much. In any game where there is anunlimited upside, but a strictly limited downside, the evidencesuggests that spread bettors tend to bet in the hope (evenexpectation) that the numbers will exceed rather than go short ofthe spread. For instance, if the spread about the number of pointsin a rugby match is 40 – 43, then the maximum downside of a buy at43 is 43 times your stake. The maximum downside of a sell is inprinciple unlimited, however, and in practice potentially verypainful. If this theory is correct, then bettors will tend to buy ratherthan sell the points, buy rather than sell the number of goals in amatch. Partly this may be to reduce the risk involved in the unlimiteddownside of a sell that goes pear-shaped. Partly, though, it may bebecause a significant element of those who bet on the matches wantto combine winning with enjoying. There is surely more pleasurein urging on the sides to score, or to get stuck in, than flinching atevery sight of player near goal or even player near player. As bettorsin the conventional spread markets have become more sophisticatedhowever, this pattern has no longer tended to hold the force of old.

The big event, like a World Cup Final or Super Bowl is, however,not just another bookings play, or Nationwide League total goalstrade. Many of the traders, beer in hand, will be betting to liven upthe game, particularly in running. Such an approach normally goeshand in hand with a buy strategy. This is the time, at least as a long-term strategy, for the more sober judges to sell – to sell supremacyand to sell points.

160

BETTING TO WIN

In other words, wait till the crowd gets hot. Now is the time toplay it cool.

161

DRAW BIASES

Chapter 3: Exploiting Disparities

On Tuesday, 7 November 2000, it was 4 pm in the UK, and the pollsin the US Presidential election had recently opened on the WestCoast of the United States. At that time, IG Index were quoting thePresidential election as follows:

Bush: 265 – 275 electoral votes; Gore: 265 – 275 electoral votes.270 votes were required for victory.At the same time, Victor Chandler were offering Bush at 8 to 15

and Gore at 11 to 8. The question is – could they both be right?At first sight, the answer must surely be no. After all, the mid-

point of the IG Index quotes are 270 apiece, i.e. level-pegging, whileChandler has Bush as almost a 2 in 3 chance.

In fact, the answer is yes, at least in theory. Let me explain. Imaginethat 260 votes are in the bag for Gore and 250 for Bush. One state isoutstanding, with about 30 votes.

Bush is about a 1 to 2 shot to win this state, and so the best estimate,mathematically, of the number of extra votes he will get is 20, i.e. a 2in 3 chance of getting the 30. Gore is about a 2 to 1 shot to win thestate, and so the best estimate of the number of votes he will obtainis 10, i.e. a 1 in 3 chance of getting 30.

Chandler’s odds, allowing for the profit margin, are close enoughto reflecting these probabilities, and so the odds add up, on thisscenario, to an expected total of 270 apiece. Sounds incredible, butnot half as incredible as what actually happened on election night!

So much for the mathematics. In reality, of course, Gore and Bushdidn’t have well over 200 electoral votes in the bag, and the realreason for the difference between the odds on offer is probably rathermore straightforward. Either there was a significant difference ofopinion between market-makers at IG and Chandler, or else someonewas a little slow to react to unfolding events. Bettors who were alertto this market imperfection could play these competitors off againsteach other, to their significant advantage.

These imperfections are growing all the time, and the Chandler/IG disparity represents just one more example of the many andvaried opportunities on offer to the discerning investor looking tomake a profit with limited risk.

162

BETTING TO WIN

The best genuine value, of course, is usually to be found where itis most difficult to find. It takes time, and a little effort, but that’swhy the rewards are there. Invest the energy, and the value can bewonderful.

The Iowa Electronic Markets web-site (www.biz.uiowa.edu/iem)is an example. You can bet (trade) on all sorts of markets, in a waysimilar to our spread markets. During the last US Presidentialelection, for example, those wishing to favour one or other of thePresidential candidates could bet with a myriad of fixed-odds outlets,on the spreads with IG Index, Spreadex and Cantor Index, or at theIowa Electronic Market. And it’s the latter option which offered thereal value.

A little close inspection was all that was required for a handsomedividend, to low risk. For while Spreadex were quoting Bushsupremacy over Gore (in terms of the popular vote) at only 0 – 0.6%,and Cantor at 0.5 – 1.1%, the Iowa Electronic Market had Bushwinning handsomely in terms of votes cast. In the event, Gore woncomfortably in terms of the popular vote, but whatever hadhappened, a little judicious staking across national borders was areal golden path to Shangri La.

163

DRAW BIASES

Chapter 4: To Arb Or Not To Arb: That

Is The Question

You flick to teletext early on a Friday morning, and turn straight tothe quotes on the match between Everton and Liverpool. Let’s saythat IG Index are offering a spread of 40 – 44 about the number ofbookings in the match. For each yellow card 10 points are awarded,and 25 for a red. On that basis, IG’s market-makers are expecting,in a statistical sense, slightly more than four yellow cards to be issuedin this match. That means that you can sell bookings at 40 and beassured of a profit if there are three yellow cards (or fewer) and nored, or buy at 44, and earn a profit if the match produces at least theequivalent of five cautions. The problem arises when you sell, andthe cards fly, or when you buy and the referee has an episode ofunnatural leniency. If only you could win either way. Sometimesyou can.

The opportunity to win either way occurs whenever one of themarket-makers sets a quote totally outside the quotes of another.Let’s say, for example, that Sporting Index set a quote for the samematch of 45 – 49 (IG, remember, are offering 40 – 44). In that case,you can sell bookings at 45 with Sporting, and buy simultaneouslywith IG at 44. Now whatever happens in the match, you areguaranteed a one point profit. To illustrate how this works, considerwhat happens if six yellows are issued and one red. The mark-up(final total) is 60 plus 25, i.e. 85. This would not be good news if youhad only sold at 45 with Sporting, as you would be facing a loss of40 points (85 minus 45). Fortunately, you win 41 points from IG (85minus 44). The net gain is one point. Consider the alternativescenario in which only one yellow card is issued. Now, buyers ofbookings at 44 with IG face a 34 point loss (44 minus 10), but sellerswith Sporting earn a 35 point profit. Once more, the net gain is onepoint.

This riskless profit, earned by simply buying and selling withdifferent companies is known as an arbitrage (or, more commonly asan ‘arb’). It occurs most usually in the bookings market, the shirtsmarket, and sometimes in the time of goal markets. Much less

164

BETTING TO WIN

common is to spot an arbitrage in goal supremacy or total goals.The question is what to do when you turn to teletext or the Internet,or your newspaper, and spot this glaring ‘arb’, with minutes still tothe opening of trading.

A lot depends on the company responsible for the arbitrage, andsometimes on your trading record with them. Sporting Index, forexample, have usually been willing to maintain the price they setbefore the opening of trading, at least for a short period, to smallstakes, but only to traders adjudged by them not to have overstayedtheir welcome in this particular field of operations. City Indexadopted a very different policy when they operated as a separatecompany in the sports spread markets, tending to shift the line at‘warp speed’ whenever their quote was out of line with the othercompanies. This ensured that you were unlikely in reality to takeadvantage of any apparent arbitrage position.

In other words, different companies have different policies, andonly with experience can you ascertain their current policy.

The bottom line is that arbitrage positions do persist, which in away is a healthy sign, inasmuch as it indicates that the spreadcompanies are not explicitly colluding on price. On the other hand,the position at advertised prices does not last long, if at all, and youcan never be sure of obtaining the quote you want to the stakes youwant. At best, you will normally earn a small return, for the risk ofnot being able in any individual circumstance to close two positionsat a prevailing arbitrage position. It is perhaps better to view anarbitrage position as an offer of good value by at least one of thecompanies. If this fits in with your own perceptions of the valueoption, consider taking it before it disappears. After all, genuinevalue rarely lasts long. Beware, though, of earning a reputation asan ‘arb shark’. You may find yourself shunned.

165

DRAW BIASES

Chapter 5: Beware The Downside!

Imagine that you are on a hiking expedition in the Americanwilderness, miles from base, and low on supplies. Suddenly you comeacross a bridge that crosses a chasm and which provides a neat short-cut home. You assess the risks, and counter each one by the correctuse of the latest safeguards available to the sensible trekker. So goesthe story told by portfolio managers in a 1995 Wells Fargo–Nikkopublication called Global Currents. The tale ends (and more importantly,so do you) as you successfully cross to the other side, only to comeface to face with a ravenous mountain lion. The moral of the story forinvestors is clear. Risk exists because more things can happen thando happen, and that encompasses far more than you might ever haveimagined, let alone quantified and analysed.

You are not likely to come across too many hungry killers at yourlocal racecourse, of course, at least not of the traditional variety, butthe principle remains, and it is why you should always be willing toaccept the maximum possible downside, regardless of howimprobable it might appear.

Fortunately, conventional betting does offer a limited downside,i.e. the stake money. If your sure thing comes unstuck, even at thatvery generous 1 to 10, at least you won’t be asked for a furtheradvance on what you’ve already staked. Not so with spread betting,where the potential for disaster can be potentially hazardous to yourwealth, if not your health.

Conventional wisdom dictates that in spread markets the bestpolicy for the cautious is to buy rather than to sell. In many cases,this is true. In the market for bookings, for example, the line mightbe set at 42 – 46. If you buy at 46, the maximum loss you face is 46times your stakes, and that will happen if there are no cards issuedat all. If, on the other hand, you sell at 42, the potential loss could becatastrophic. Picture the scene: ten minutes left, and you are sittingpretty without so much as a caution in sight. Then a fight breaksout, and suddenly red cards are flashing everywhere. Truly a caseof the agony outweighing any previous ecstasy. Be careful ofgeneralisations, however, for in certain markets the buy may carry aquite unforeseen risk .

166

BETTING TO WIN

A classic case in point is the traditional mini-performance markets.One method of scoring these is to award 15 points for a win, 10 fora goal, 5 for a draw, and 5 for a clean sheet. At 32 – 36, you mightthink that a buy could in the worst scenario ‘only’ damage you tothe tune of 36 times your stake, while just consider the upside inthat dream eight-goal walkover. In the small print lies the ‘lion’,however, and that is the deduction of 15 points per red card. All ofa sudden that ‘maximum’ downside is not such a maximum afterall!

If you must know your total potential liabilities in advance, oneway is to stick to buying goals. You know for sure that they can’t fallbelow zero. You might even sell or buy minutes to the first goal.This might be one of the more unpredictable markets, but at leastyou know that no team can score in less than no minutes, or after 90minutes (stoppage time is ignored for betting purposes).

Alternatively, you can enter the markets that offer a maximumstop-loss. This acts as a kind of insurance policy to limit your losseson any one event, such as a trade on the total lengths by which onehorse beats another. This can be quite a comfort when your horsetrips over its own feet coming out of the starting stalls. You can alsoagree a stop-loss in advance in certain defined circumstances .

Another option is to turn to the fixed-odds equivalent of spreadmarkets. So instead of buying corners at 15, you could place a fixed-odds bet on there being more than 15 corners at, say, 8 to 1. Insteadof buying bookings, you could place a fixed-odds bet on more thaneight yellow cards being issued at, say, 12 to 1. This approachcertainly limits your risk, but the in-built margins associated withthese speciality markets are somewhat off-putting, not to saydestructive of value.

In the end, then, there really is no such thing as risk unless it iscoupled with a consequence, and that can be different for differentpeople. Even so, you might do worse than always to bear in mindthe famous ‘law’ that states that whatever can go wrong will gowrong. It might one day keep the shirt on your back, and that’s forstarters.

167

DRAW BIASES

Chapter 6: Quarb Strategy: Using

The Odds To Beat The Odds

The date is Sunday, 24 March 2002. Liverpool are playing Chelsea.The four major spread betting companies are offering a variety ofodds about the likely number of bookings in the match. IG Indexare offering a quote of 46 – 50. One way of looking at this is thatthey expect there to be a little under five yellow cards, statisticallyspeaking, which would produce a make-up of 48 bookings points.This is the mid-point of their quote, and their margin lies in the sizeof the spread around this point. Cantor again set a quote of 46 – 50,as do Spreadex. Of course, the actual outcome cannot be 48 exactly,since yellow cards pick up ten points, and red 25. Instead, the 48 isa statistical estimate of the expected value. Essentially, these market-makers are estimating that there will be about four or five cautions,but that it is rather more likely to be five than four. Sporting Indexare a little out of line, offering a quote of 50 – 54. They are acting,therefore, as what I call the ‘maverick’ bookmaker.

Now let’s say that you know no more about the likely number ofbookings in this match than you do about the finer points ofEinstein’s General Theory of Relativity (any theoretical physicistsreading this excepted). Can you still make a meaningful trade inthis market? The answer is yes, so long as you use the odds againstthemselves. The method works like this. Take the mid-point of eachquote in turn, and average them. So for IG the mid-point is 48 (halfway between 46 and 50). Now do the same for Cantor and Spreadex,and on each occasion you again get 48. For Sporting, however, themid-point is 52. Add these up and you get a total of 196 (48 plus 48plus 48 plus 52). Now divide by four, and you arrive at the averageexpected bookings make-up based on the quotes of all the market-makers. In this case, the answer is 49 (196 divided by four). This iswhat I call the golden number. The beauty of this approach is thatyou have used the combined wisdom and experience of fourprofessional market-makers to calculate the most likely outcome ofthis event.

The question is whether this can be turned to your advantage in

168

BETTING TO WIN

a practical sense. The answer is yes, in certain circumstances. Thecircumstances are where one of the quotes falls everywhere outsidethe golden number. In this example, it is 49 and Sporting Index’squote of 50 – 54 lies everywhere outside it. All the other quotesinclude the golden number. If we assume that the market-makers,taken as a whole, are on average correct, then a sell of bookingspoints at 50 with Sporting Index should yield an expected profit of1 point. This is what I call a Quarb, and it relies on the assumptionthat the market as a whole generates a better forecast of the bookingsmarket than the outlying market-maker. If so, you can make a long-term profit while knowing next to nothing about the teams’disciplinary records, or that of the referee.

Does Quarb trading really work? In theory, finding the answersounds as easy as collecting the data and testing. In reality, there area number of problems to resolve. For example, what happens whenthere is an outright arbitrage position, i.e. where one of the quoteslies totally outside that of another? An example would occur ifSporting Index quoted 2.7 – 3.0, but IG Index, say, offered 2.3 – 2.6.Clearly here you could buy at 2.6 (with IG) and sell at 2.7 (withSporting) and win whatever the result. Should such circumstancesbe defined as Quarb positions also? If so, might this not skew theresults, particularly if it is not possible in practice to trade at theseprices? On the other hand, is it justifiable simply to ignore them?

Again, what happens if the top end of one quote is short of theaverage mid-point by exactly the same amount as the bottom end ofanother quote is above it? Which, if either, do you choose as the‘maverick’ position?

And even if you resolve these issues, on what basis do you decidethat any profits made in retrospect from trading according to sucha strategy are more significant than chance?

In a paper I presented, with my colleague David Paton, at the2002 Royal Economic Society conference at the University ofWarwick, a start was made in tackling these issues on the basis of asignificant sample of matches.

To be precise, spreads for the bookings market were collected forup to five companies that existed during the period of theinvestigation. The opening prices are usually announced one or twodays before each game, and were collected throughout the two

169

QUARB STRATEGY: USING THE ODDS TO BEAT THE ODDS

seasons. Data were collected about 207 matches for the 1999/2000season and 240 matches for the 2000/2001 season.

In the 1999/2000 season, the Quarb strategy suggested 60 trades.On the basis of £1 being placed on each Quarb trade, a bettor wouldhave won £973 on 36 winning bets and lost £384 on 22 losing bets, anet profit of £589 over the season. The remaining two bets yielded areturn of zero.

The 2000/2001 season, 80 trades were suggested by the Quarbstrategy. The average return to these bets was lower than in 1999/2000 but would still have yielded winnings of £836 on 50 winningbets and losses of £450 on 28 losing bets, a net profit of £386.

The conclusion of the study was that the mid-point of all thequotes on offer is a better forecast of the actual outcome in thebookings market than is the mid-point of the spread offered by themarket outlier (the ‘maverick’). This casts doubt on a hypothesisthat market-makers who set quotes out of line with the prevailingview do so because they possess better (even privileged) information,or that they are able to process a given set of information moreeffectively than the market as a whole.

Moreover, using the notion of quasi-arbitrages or Quarbs, it waspossible to devise a trading strategy on the basis of the outlyingspread that yielded significantly positive profits.

170

BETTING TO WIN

Chapter 7: Tomorrow’s Another Day

It is said that on returning from a day at the races, a certain LordFalmouth, of yesteryear, was asked by a friend how he had fared.‘I’m quits on the day,’ came the triumphant reply. ‘You mean bythat,’ asked the friend, ‘that you are glad when you are quits?’ Whenhis Lordship replied that indeed he was, his companion suggestedthat there was a far easier way of breaking even, and without thetrouble or annoyance. ‘By not betting at all!’

The noble Lord said that he had never looked at it like that beforeand, according to legend, gave up betting from that very moment.

We all know the feeling, but it doesn’t help us at all when that betwe were so sure about, and so nearly placed, goes in at 20 to 1. Ifonly! Of course, you have not actually lost any money; you are noworse off than before. Still, that doesn’t stop you kicking yourselfover the £200 or so that you would have won.

This is an example of what economists call ‘regret theory’ in action.According to the theory, what motivates us is not just what we have,but also what we might have had if only we had behaved differently.The theory might even explain in part why bettors at the racetracktend to bet too much on longshots. Just look at how much you mighthave won!

Apply this logic now to the case of a football bet in running. Beforethe game started, you really were going to buy goals in that scrap.After all, with two of the leakiest defences in the League playingeach other, there must be goals, and lots of them. You pick up the‘phone, get out the first syllable – and then remember that bill whichneeds paying. You replace the receiver.

After five minutes, an ill-judged tackle inside the penalty arealater, and it’s 1 – 0, with all of 85 minutes to play – all of 85 minutesto regret that moment of indecision, all of 85 minutes to watch thegoals pile up. You’d better pick up the blower and buy before there’sa second. Hold on a moment, though, the quote has gone up a wholegoal, so you are effectively back to where you started. If you buynow, there could be three goals, and you’re still in loss.

This is where regret theory gets complex, which is why now isthe time to apply the cold rules of that old fallback, statistics and

171

TOMORROW’S ANOTHER DAY

probability theory. Conventional theory suggests that the numberof goals in a football match shoud be distributed, on the average, ina timeless fashion, called a Poisson distribution. This allows us tocalculate how soon, on the average, the first or second goal of amatch will be scored if we have a good estimate of the number ofgoals that will be scored. It also allows us to estimate how muchsooner one team will score, given the estimated superiority of oneteam over another.

There is also evidence that goals are somewhat clustered together.In other words, goals beget goals. If one goal is scored, it is just thatlittle bit more likely that another is just around the corner. Talking ofcorners, there is evidence that the same pattern applies to these aswell. Just like waiting for a bus, there may be nothing for a whileand then two or three all at once.

Apply this cold logic now to your quandary as you sit, telephonein hand, cursing your luck for not buying goals earlier. The reality isthat the early goal has made another strike just that bit more likely,and unless the quote goes up by more than a goal, the margin is inthe buy. In practice, however, the margin is usually swallowed upsomewhere between the two ends of the quote.

The moral of the tale is in what it suggests you should not do.You should not, at least on the whole, sell goals just after a goal, atleast not just because there has been a goal. Unlike lightning, strikersdo strike twice, and the evidence is the same even if that fifth minutegoal turns out on this occasion to be the only one in the game.

Of course, you may live to regret your decision, but given thechoice between potential regret and actual probabilities, it is withthe laws of probability that I for one side would always side.

172

BETTING TO WIN

Chapter 8: Betting On The Election

Spreads

It was once said that only three people ever understood thatnineteenth-century conundrum of European diplomacy, the‘Schleswig-Holstein question’. Of these, one was dead, another hadgone mad, and the third had forgotten. Much the same has beenclaimed of aspects of quantum physics. What we do know of thisarea of human enquiry, however, can tell us much about the waysof betting markets.

In particular, one interpretation of quantum theory is that nothingactually exists for sure, only with a higher or lower degree ofprobability. In the quantum world this probability may be so high,however, that it is in fact a sure thing for all practical purposes.

What unites quantum theory with conventional probabilityanalysis is that every event is subject to random disturbances.Sometimes these are predictable and even out over time, based ontheir relative frequency. The chance of a coin turning up heads, or adie showing a six, are clear examples. The odds are readily calculable,and it is possible on this basis to determine quite easily whether theedge is in your favour or not.

Where probability analysis differs from quantum theory, in onesense at least, is when the outcome is already determined, but wedon’t yet know what it is.

It is said that that the eccentric racehorse owner of yesteryear,Dorothy Paget, had an arrangement with her bookmaker to bet ‘aftertime’ this way. Apparently, she was asleep during the time of racing,and liked to wager on the day’s events immediately after waking.

More conventionally, we are talking about the type of bet wherethe outcome is known to a very select band, and which can beparticularly profitable if you are one of this band. As a market-maker,you may be willing to lose some money to these insiders, in order toformulate a general market which can attract much more moneythan this from those who are not ‘in the know’.

It is movements in the odds in such ‘already determined’ marketswhich are worth following particularly carefully. There are

173

BETTING ON THE ELECTION SPREADS

handsome dividends to be earned from identifying these markets,and I am not just referring to recent match-fixing allegations.

Election time is one such opportunity. The reason for the value atelection time derives from the plethora of polling, private and public,which is taking place. Much of this unfolding evidence is availableto those close to the survey well before it is made public. Movementsin the spreads with no obvious source can often be traced to this,and particularly so in the closing stages of an election campaign. Itis why the markets used to shut well before close of polling.

Bookmakers are rather bolder these days, and sophisticated bettorsalert to odds movements can benefit greatly. The key lies in an alertnessof all public information, while keeping a close eye on marketmovements. Any shift with no obvious reason is the one to look outfor, for it is quite likely to presage as yet unpublished new information.Significant movements in these essentially predetermined positionsare just that – significant, in what they tell us about the thinking ofthose who are either in the know, or at the very least close to ‘theknow’.

As long as this new information is not fully incorporated into thenew odds, now is the time to act. The opportunity exists becausemost markets take time to adjust to new information, whateconomists call taking time to establish a new equilibrium. In bettingterms, this roughly translates to a variation of ‘the early bird catchesthe worm’. In the right circumstances, that particular kind of wormcan be very tasty indeed. Catch it while you can.

174

BETTING TO WIN

Chapter 9: Control The

Consequences

There really is no such thing as risk without consequence, and thekey to successful long-term betting is to keep the consequences verymuch under control. Let me explain.

In traditional fixed-odds betting, there is a great deal of evidenceto suggest that punters are what economists call ‘risk-loving.’Basically, this means that they are willing, consciously or otherwise,to sacrifice some expected return for taking on a little more risk.Most famously, this occurs when bettors at the racetrack pile on thebig prices. What is a little strange about this phenomenon is thatlongshots not only pay out worse on average than shorter-pricedanimals, but bettors are also taking on a more risky proposition inbacking them. After all, they win less often, and so any return youmay obtain is likely to occur less frequently. Statisticians say that thepunters are taking on higher ‘variance’ (risk) for a lower ‘mean’(average) expected return. In other words, these bettors are eitherrisk-loving, or at least they are behaving as if they are.

Spread bettors are a breed apart, however, to judge by theirbehaviour in their own chosen betting forum. Unlike traditionalpunters, they behave more like financial investors, who are generallyacknowledged to be ‘risk-averse’, i.e. they will require a higherexpected return to compensate for any extra risk. The clearestevidence of this can be found in the quotes offered about the timeof the first goal in any football match. In a standard match, with anexpected goal tally of, say, 2.6 goals, one would expect that the firstgoal would be scored in about the 36th minute. The problem withthis market for risk-averse traders is that the downside to a sell ofthe first goal time is quite sizeable, i.e. 54 (90 minus 36) times thestake. This is also a very volatile market, where the actual outcomesare spread widely around the quoted position. In this sense,therefore, a trade in this market, and particularly a sell, is extremelyrisky. As a result, at the ‘correct’ quote, buy trades are likely tooutweigh sell trades. If this hypothesis is correct, we would expectto see the mid-point of the quotes on offer about the time of the first

175

CONTROL THE CONSEQUENCES

goal in a match set on average a little above the level determined byobjective calculation. And we do!

The same applies, in spades, in the quotes on offer about the timeof the first goal in any match in a tournament (in seconds). Justimagine a sell of seconds at, say, 110 seconds (the standard quote atthe opening of World Cup 2002), for a fiver a second. What if thefirst goal in any tournament match is not scored until the 360thsecond? It sounds a perfectly reasonable scenario, but that means aloss of (360 minus 110) times 5, or £1250. The maximum gain is, ofcourse, just 109 times your stake, even if a goal is scored in the firstsecond.

Because of this, most bettors in this market tend to err on the sideof caution and buy, which means that the quote is usually ratherhigher than it should be.

The World Cup was in fact a classic example, where clued-upsellers needed a nerve of steel to pick up the jackpot when Turkeyscored in the 10th second against South Korea, in the penultimategame of the tournament. Until then, a stiff loss beckoned. Fancyselling seconds to the first goal next time?

A sell of runs in a cricket match is another obvious example. Anextra ingredient here is the ‘fun factor ’, though it only applies inparticular circumstances. To illustrate, there is surely an extra touchof spice in a buy of runs in a cricket match compared to a buy ofminutes to the first goal in a football match. After all, while the firstis consistent with action and flair, the second is associated withrecoiling every time the ball is played anywhere near the strike zone.

All in all, though, a sober-minded risk-averse trader will simplyseek to avoid the twin devils of significant liabilities and significantvolatility. Adding together these two factors suggests that the averagetrade will be skewed a little too heavily towards the buy side of themarket as compared with an objective assessment of the probabilities.This means that the quote will tend to be slightly higher than itshould be. This should occur in any event where the upside to abuy trade is greater than the potential downside.

This is not to say that the path to long-term prosperity is througha simple sell strategy in these markets, for while the quote may be alittle too high it may not be lifted sufficiently to cover the size of thespread.

176

BETTING TO WIN

The key to a successful betting strategy is to take account of allrelevant factors. For a football match, for example, identify caseswhere you genuinely think that the first goal will be scored morequickly than implied in the quote, or where you expect the numberof runs in that innings to be on the low side. With intuition workingin concert with analysis, now is the time to act.

But what about the risk? Well, risk as I stated earlier is nothingwithout consequence, and the smaller the stake, the smaller theconsequence. Controlled losses, and odds in your favour, really dotogether add up to time on your side. And money too!

Part Five: Beating The Exchanges

Chapter 1: Backing And Laying On

The Exchanges

Betfair operates by allowing clients to either back or lay a bet. Inorder to open an account as a client, a deposit must be made. Creditcard deposits carry a 1.5% charge, but there is no charge for UKdebit card deposits or withdrawals to any card. Deposits andwithdrawals can also be made by cheque and bank transfer, thoughthere is a charge for bank transfer withdrawals.

When you bet on an event in the conventional sense, this is termedbacking the event. If you wish to act as a bookmaker and offer oddsabout an event happening this is known as laying the bet.

Betfair provide the best three offers on each side of the equation.Thus the best three odds at which you can back an event aredisplayed on the left-hand side of the screen, together with theamounts that you can bet at those odds. On the right-hand side ofthe screen are the best three odds at which you can lay an event.

Betdaq, GGBet and Sporting Options also operate by allowingclients to either back or lay a bet. In the case of GGBet, however, theterminology is different. If you wish to back an event happening(i.e. to bet on something in the conventional sense) their term isthat you take the odds.

Betfair, Betdaq and GGBet all provide the best three offers on eachside of the equation. Thus the best three odds at which you cantake the odds (back) are displayed on the left-hand side of the screen,together with the amounts that you can bet at those odds. On theright-hand side of the screen are the best three odds at which youcan lay a bet. In addition to odds betting Betfair also offer line bettingand range betting. Let’s consider each of the three types of bettingin turn.

180

BETTING TO WIN

Chapter 2: Odds Betting

IntroductionWith odds betting the bettor predicts whether an outcome will occuror not, and places a bet on that outcome at given odds. Inconventional betting markets, bettors place their bet at odds set bybookmakers, i.e. bettors take odds and bookmakers make (or offer)odds.

Betting exchanges permit clients to both take and to offer odds.In this way, you can bet both for and against an outcome occurring.When you take odds you back the outcome and when you offerodds you lay the outcome.

Clearly, if you back an outcome at given odds, you will gain ifthat outcome occurs. Say, for example, you back England to beatAustralia at cricket at odds of 3 to 1, for a stake of £10. This meansthat you make a profit of £30 if England win (3 times your £10 stake),plus your £10 stake returned. Your loss if England do not win islimited to your stake, i.e. £10.

Laying an outcome, as explained earlier, is the exact opposite ofbacking it. Essentially, it is betting that the outcome will not occur.This is what bookmakers traditionally do. For example, if you layAston Villa at odds of 4.0 for £100, you think it is sufficiently unlikelythat Villa will win that you are willing to offer the equivalent oftraditional odds of 3 to 1 against it. Your total return, if Aston Villalose or draw, is £100 (i.e. the stake), and your maximum loss is £300.In effect you are accepting the bets of other clients who believe that3 to 1 is worth taking about Villa’s chances of winning.

What Are Digital Odds?Digital odds differ from the traditional fractional odds in that theyinclude the stake. For example, if you place a bet of £100 at digitalodds of 5.0, then the return to a winning bet is £500 (which includesthe £100 stake). The fractional odds equivalent is 4 to 1.

What If The Odds You Want Aren’t Available?If you want to back an event happening at better odds than are

181

ODDS BET TING

currently available, you are able to place an order for a better price.In doing so, you are hoping that somebody will see your offer andbe willing to lay you the bet, partially or in total, at the odds yourequest.

Your offer to back appears on the right-hand side as a request forsomeone to lay you a bet, at the odds requested and up to the amountrequested.

ExampleFor a few minutes after opening on Sunday 26 May 2002, it was

possible to obtain 7 to 1 in a string of betting offices, about QuarterMoon winning the Irish 1000 Guineas. Your stake size would havebeen restricted, but say you placed £100 at that 7 to 1:

Returning to your computer screen you would have noted thefollowing Betfair odds, available at about 12.30 pm that day:

Available to back: At 6.4 (£382); At 6.6 (£67); At 6.8 (£25)Available to lay: At 7.0 (£873); At 7.4 (£145); At 7.6 (£132)

You could now lay Quarter Moon at 7.0 (or 6 to 1) for £100.If Quarter Moon wins, your profit at the betting office is £700, but

you only pay out £600 to your betting exchange colleagues.If Quarter Moon loses, you lose £100 at the betting office, but you

win £100 (minus commission) in the betting exchange. In otherwords, you stand to win a net profit of £100, for the maximumdownside of 5% of £100, i.e. £5.

By laying the horse at 6 to 1 for £110, say, you would make a profitof £90 if Quarter Moon wins, but make a small profit, even net ofcommission, if it loses.

Now let us see how exactly you would go about laying the horseat 6 to 1 on Betfair. First, you select Today’s Horse Racing from thevertical list of sports/special bets available on the left side of thescreen.

You then select 4.00 Curragh from the list of options presented,and you are led to a list of the best three prices at which it is possibleto back and lay the horse, together with the amounts at which youcan back or lay at these odds (see above).

As noted in the example, it was possible to lay Quarter Moon at

182

BETTING TO WIN

7.0 up to a total of £873. Note that this means that you are offeringodds of 7.0 (6 to 1) to other clients of the exchange who wish to backthe horse at that price. If you agree to accommodate all this money,your net liability is, therefore, more than £5000. To be precise, it is 6× £873, or £5238, which means that you would need to have thatamount of money deposited in your account, over and above anyoutstanding potential liabilities on other events you may alreadyhave laid.

Let us say that you decide to lay Quarter Moon at 7.0 for £100. Todo this you select the LAY button, and click on it.

If you wish to lay at 7.0 for £100, simply enter £100 at this point.You are immediately warned of the implications of what you are

intending to do, i.e. that you stand to win £100 (minus commission)if the horse does not win, but to lose a net amount of £600 if thehorse does win.

Press Enter again to confirm.An alternative approach is to offer to lay the horse at less than 7.0,

say at 6.8 when the display prompts you with the possibility ofchoosing your own odds. You now select 6.8 for £100, say, and pressEnter. The implications of your proposed offer are again displayedbefore you finally confirm.

If Quarter Moon does not win, you keep your stake of £100. Intotal your gross profit is £100 (minus commission). If Quarter Moonwins, your gross payout is £680. This includes your stake of £100. Intotal, your loss is £580.

If you confirm, an extra £100 is added to the BACK side of thescreen, which means that those wishing to back the horse at 6.8 (i.e.5.8 to 1) will have the opportunity of staking £100 with you, inaddition to however much is being offered by other members of theexchange at those odds.

In the example above, £25 was available at 6.8. With your offer,£125 is now available. Of course, this is only an offer, and there is noguarantee that anyone will be willing to take your offer at thoseodds. In the meantime, you may have missed the opportunity toLAY the horse, for a sure profit, at 7.0 when people were willing toback it at that price.

183

ODDS BET TING

When Should You Back Or Lay?This depends on the type of bettor you are. If you are totally averseto taking any risk, you will only lay when you can do so at a shorterprice than you have taken. In other words, you will only lay at 6 to1 if you have previously obtained longer odds than 6 to 1, or arecertain to obtain longer than 6 to 1 at some future time.

If you are willing to gamble, however, you might do worse thantake a look at the following strategy:

1. Identify cases where the back and lay prices are quite closeto each other. This way, the price you are getting will bequite close to what the market expects. For example, if youcan back a horse at 6.8, or lay it at 7.0, whichever you do willprobably be close to the true probability. Moreover, there isa reasonable chance that the market will move enough toallow you to back and lay for a sure profit. After all, it doesn’thave to move far.

2. Take a look at the relative amounts of money on both sidesof the screen. In the example just given, £25 was available toback Quarter Moon at 6.8, but all of £873 to lay it at 7.0. Thismeans that clients were willing to stake up to £873 if youwere willing to offer them odds of 7.0 (6 to 1). That’s a signof confidence in the horse at that price, just as a queue at thecake stall is a sign of confidence in the cakes. On the otherhand, a paltry £25 to back at 6.8 means that clients are onlywilling to let you back the horse at that price for £25. Soundslike the market is not too confident in letting backers takethat price.

If you know nothing else, in these circumstances a punt at the 6.8looks the better option. At those odds, we are still only saying thatthe horse has about a 1 in 7 chance of actually winning the race, butyou might be on the right side of taking the value.

For the record, Quarter Moon, trained by Aidan O’Brien andridden by Mick Kinane, came second.

184

BETTING TO WIN

Chapter 3: Line And Range Betting

In addition to standard exchange betting, Betfair also offer line andrange betting on certain events.

What Is Line Betting?Line betting is an even-money bet on some numerical event, suchas the number of runs in a cricket match, corners in a soccer match,minutes to the first goal, and so forth. The line in question is a lineof numbers representing all possible results.

A bet is undertaken by a decision to either buy or sell the line, as itis termed, at a given price. A buy of the line is rewarded if theoutcome is above (greater than) the line, and a sell of the line isrewarded if the outcome is below (less than) the line.

For example if the line on England first innings runs in the firstTest of the Ashes series is set at 220, then a buy of the line would bea winning bet (at evens, i.e. 2.0) if England score more than 220runs. Thus, a buy of the line at 220 for, say, £50 would return a netprofit of £50 on the bet. If England scored less than 220 runs, on theother hand, you would lose your £50.

Similarly, if you sell the line at 220 runs, and England score fewerthan 220 runs, you win a net profit equal to your stake. If you sell at220 and England score more than 220 runs, however, you lose yourstake money.

If England score exactly 220 runs, the bet is void, and you winand lose nothing.

The advantage of line betting over conventional spread betting isthat your losses are capped at the stake. The disadvantage is thatyour profits are capped as well.

The line is not always set at a whole number. For example, Englandruns might have been set at 219.75. The way to look at this is to seeit as two bets, one at 219.5 and the other at 220. If your stake is, say,£200, this can therefore be viewed as two bets of £100 each. If youbuy the line in this way at 219.75, and England score 220, one half ofthe bet is a winner (the £100 buy at 219.5) and the other half is void(the £100 At 220). In other words, your net profit for a £200 stake is£100.

185

LINE AND RANGE BETTING

What Is Range Betting?Range betting is effectively betting on some numerical event, suchas the number of runs in a cricket match, corners in a soccer match,minutes to the first goal, and so forth, around a predetermined‘points index’. The effect of range betting is that the bettor earns aprofit or incurs a loss equal to the difference between the level (orprice) at which the bet is placed and the eventual outcome,multiplied by the unit stake, as specified by the bettor. The ‘stake’ isdefined as the amount bet per unit, e.g. per run or per point.

In range betting, a bet is undertaken by a decision to either buyor sell at a given level, or price. A buy is rewarded if the outcome isabove (greater than) the level at which the buy is implemented, inproportion to the amount by which it is greater. A sell is rewarded ifthe outcome is below the level at which the sell is implemented, inproportion to the amount by which it is less.

For example, if you buy England runs at 220, for £10 per run,then you win £10 for every run by which England exceed this total.On the other hand, if England score fewer than 220 runs, you lose£10 for every run by which the total falls short.

The price need not be a whole number. For example, the priceoffered about the total number of goals scored in a soccer matchtypically lies between 2 and 3, say 2.6. Assume that you buy totalgoals at 2.6, then whatever happens you must either win or lose. Itis not possible to break even, for the simple reason that there cannotbe exactly 2.6 goals in practice. For example, if you buy total goals at2.6 for £50 per goal, then you will win if there are three goals ormore, and lose if there are two goals or less. If there are four goals,for example, then you would win £70, i.e. £50 × (4 - 2.6). If only onegoal was scored, however, you would lose £80, i.e. £50 × (2.6 - 1).

Unlike in spread betting markets, Betfair impose a minimum andmaximum limit on range bets. In a range bet for corners, for example,the range is between 7 and 17 corners.

Cancelling A BetAny offer (whether to back or lay) that has not been matched can becancelled. It is not possible however, to cancel a bet once it has beenmatched, i.e. the offer has been accepted by another client.

186

BETTING TO WIN

Chapter 4: Asian Handicap Betting

On The Exchanges

Asian handicap betting in a betting exchange environment operatesin the same way as odds betting, except that an artificial handicap isgiven to one of the teams to bring the odds about each side winningcloser to each other than would otherwise be the case.

Asian handicap betting has been explained in Chapter 1 of Part 1but an example from a real match on Betfair is used here to explainexactly how it works within an exchange environment.

England v. Sweden (World Cup, 2002)2 June, 2002 – 10 am (kick-off 10.30 am)

Available to back Available to lay

England - 0.5 & - 1.0 – – 3 – –

(£75)

Sweden + 0.5 & 1.0 – – – 1.76 –

(£50)

England - 0.5 2.14 2.16 2.18 2.24 2.26 2.5

(£250) (£483) (£1000) (£800) (£10) (£20)

Sweden + 0.5 1.64 1.76 1.8 1.86 1.88 1.94

(£20) (£62) (£700) (£50) (£350) (£685)

England 0 & - 0.5 1.8 1.82 1.84 1.9 1.92 1.94

(£250) (£760) (£112) (£274) (£205) (£113)

Sweden 0 & + 0.5 2.1 2.38 2.4 2.5

(£290) (£200) (£550) (£50)

England 0 1.5 1.52 1.58 1.8 1.84 2

(£25) (£42) (£100) (£20) (£50) (£2)

Sweden 0 2.26 2.96 2.98 3.32

(£20) (£55) (£102) (£20)

Here we see that England are favourites, at scratch handicap(available to back at 1.58 compared to 2.26), but that Sweden arefavourites given a half goal start (available to back at 1.8).

187

ASIAN HANDICAP BETTING ON THE EXCHANGES

Backing England At ScratchLet’s say you did back England at scratch handicap (no handicap);let’s see what would have happened.

Say you staked £10 at 1.58. If England won, you would earn aprofit (ignoring commission) of £5.80. If Sweden won, you wouldlose your £10 stake. In the event, the match ended as a draw, and soyou would have won and lost nothing. Your stake is returned.

Backing England With A Half Goal DeficitLet’s say you backed England at - 0.5 (a half goal handicap deficit);let’s see what would have happened.

Say you staked £10 at 2.18. If England won, you would earn aprofit (ignoring commission) of £11.80. The actual 1 – 1 draw meantthat England lost by half a goal (their handicap), and so you loseyour stake.

Similarly, a bet at 1.8 about Sweden (given a half goal start) meansthat a draw becomes a winning bet, of an additional £8 to add toyour £10 stake (minus commission).

To summarise:

England - 0.5 v. Sweden + 0.5Back England: You win if England win. You lose if England draw,

or Sweden win the match.Lay England: You win if the match is a draw, or Sweden win. You

lose if England win.Back Sweden: You win if the match is a draw, or Sweden win.

You lose if England win.Lay Sweden: You win if England win. You lose if the match is a

draw, or Sweden win the match.

England 0 v. Sweden 0Back England: You win if England win. If the match is a draw,

your stake is returned. You lose if Sweden win.Lay England: You win if Sweden win. If the match is a draw,

your stake is returned. You lose if England win.Back Sweden: You win if Sweden win. If the match is a draw,

188

BETTING TO WIN

your stake is returned. You lose if England win.Lay Sweden: You win if England win. If the match is a draw,

your stake is returned. You lose if Sweden win.The same principles apply regardless of the size of the handicap.

Take, for example, the game between Brazil and China, played onSunday, 8 June. Here the Asian handicap lines stretched to 2 goals,and it is one of the more complex possibilities. If you can masterthis, you can master any of the possibilities. Here it is, with asummary of the outcomes given all possible eventualities:

Brazil - 1.5 & - 2.0 China + 1.5 & + 2.0Back Brazil: You win if Brazil win by three or more goals. If Brazil

win by two goals, you win half your bet. You lose if China win, thematch is a draw, or Brazil win by one goal.

Lay Brazil: You win if China win, the match is a draw or Brazilwin by one goal. If Brazil win by two goals, you lose half your bet.You lose if Brazil win the match by three or more goals.

Back China: You win if China win, the match is a draw or Brazilwin by one goal. If Brazil win by two goals, you lose half your bet.You lose if Brazil win by three or more goals.

Lay China: You win if Brazil win by three or more goals. If Brazilwin by two goals, you win half your bet. You lose if China win, thematch is a draw or Brazil win by one goal.

Part Six: World Cup Fever

Chapter 1: Different Betting

Strategies

In the last chapter, we looked at Asian handicap betting on theexchanges, using England’s opening match against Sweden as anillustration of how it works.

In this final part of the book, I will use the same match to highlightsome of the different ways in which a betting strategy can beimplemented.

Fixed OddsWe begin on the Saturday evening preceding England’s opener. Theaverage punter would go to his local bookie and perhaps take theodds on offer about one or other outcomes. The reality is that thechoices on offer were simply wondrous.

Take the fixed odds first, and consider the case of the typicalLadbrokes punter (I choose Ladbrokes simply as an example of oneof the big established chains). The odds about the three basicoutcomes were as follows (decimal odds in brackets):

England to win: Evens (2.0)Sweden to win: 12 to 5 (3.4)Draw: 2 to 1 (3.0)Evens is the equivalent of a 50% chance (100/2.0)12 to 5 is the equivalent of a 29.4% chance (100/3.4)2 to 1 is the equivalent of a 33.3% chance (100/3.0)

This adds up to a total implied chance for the three possibleoutcomes of 112.7%. This is known as the ‘over-round’. If the oddsreflected the true chances they would add up, of course, to 100%. Inother words, the odds were stacked in favour of Ladbrokes by 12.7%.

The more sophisticated bettor has only to look around to the maincompetitors in the market. A convenient step along this route is aclick to the Oddschecker site, and thereby to the odds available with18 other mainstream bookmakers. This revealed the following bestprices:

192

BETTING TO WIN

England to win: 5 to 4 (2.25) – with Paddy Power, NetbetsportsSweden to win: 14 to 5 (3.8) – with Totalbet, LittlewoodsDraw: 11 to 5 (3.2) – with Littlewoods5 to 4 is the equivalent of a 44.4% chance (100/2.25)14 to 5 is the equivalent of a 26.3% chance (100/3.8)11 to 5 is the equivalent of a 31.3% chance (100/3.2)

This adds up to a total implied chance for the three possibleoutcomes of 102%. This is less than a sixth of the Ladbrokes in-builtmargin (2% instead of 12.7%).

If that is not enough, there is always the possibility of a foray intothe Betbrain site, which lists odds of a much wider range ofbookmakers from across the globe, some relatively unknown. Thissite revealed the following best prices (decimal odds only):

England to win: 2.3 – with Lion BetsSweden to win: 4.35 – with SportsTabDraw: 3.35 – with SportsTab2.3 is the equivalent of a 43.5% chance (100/2.3)4.35 is the equivalent of a 23.0% chance (100/4.35)3.35 is the equivalent of a 29.9% chance (100/3.35)

This adds up to a total implied chance for the three possibleoutcomes of 96.4%. This is what is known as an under-round bookfor the market, though not for any individual bookmaker. This meansthat the margin now favours the selective bettor, who could inprinciple win whatever the result by choosing stakes on eachoutcome at the set prices, in the correct proportions.

In practice, it often means taking on trust the reputations of someof the smaller bookmakers. The odds may also have altered fromwhen you see them displayed on the odds comparison site to thepoint of actually implementing the bet. The stake size may also belimited.

Besides the traditional fixed odds outlets, the astute bettor mayalso consult the betting exchanges for alternative odds. Often theseodds will be better, since the bookmaker is cut out of the equation,but in considering this option the bettor needs to take account ofany commission charged by the exchange operator.

193

DIFFERENT BETTING STRATEGIES

Let’s look at what really was going on in the betting exchangeswhile the prices displayed above were being highlighted on the oddscomparison sites.

First, to Betfair, which was showing the following:


England 2.1 2.15 2.2 2.25 2.3 2.35

(£9169) (£3177) (£974) (£11657) (£5356) (£60)

Sweden 4.2 4.3 4.4 4.5 4.6 5.2

(£1123) (£558) (£218) (£546) (£1000) (£20)

Draw 2.9 3 3.1 3.2 3.3 3.4

(£113) (£1626) (£3951) (£5042) (£2258) (£500)

Calculating the over-round to back2.2 is the equivalent of a 45.5% chance (100/2.2)4.4 is the equivalent of a 22.7% chance (100/4.4)3.1 is the equivalent of a 32.3% chance (100/3.1)Over-round = 101.5% (ignoring commission)

Calculating the over-round to lay2.25 is the equivalent of a 44.4% chance (100/2.25)4.5 is the equivalent of a 22.2% chance (100/4.5)3.2 is the equivalent of a 31.3% chance (100/3.2)Over-round = 97.9% (ignoring commission)

From the backer’s point of view, this ‘over-round to back’ shouldbe as low as possible, and ideally below 100%.

From the layer ’s point of view, this ‘over-round to lay’ should beas great as possible, and ideally above 100%.

The odds available with Betdaq were as follows:


England 2.0 2.10 2.15 2.20 2.25 2.30

(£5010) (£452) (£261) (£265) (£400) (£13)

Sweden 4.0 4.1 4.3 4.4 4.5 4.6

(£167) (£80) (£109) (£299) (£350) (£108)

Draw 2.9 3.0 3.1 3.25 3.3 3.4

(£87) (£1125) (£145) (£338) (£368) (£480)

194

BETTING TO WIN

Calculating the over-round to back2.15 is the equivalent of a 46.5% chance (100/2.15)4.3 is the equivalent of a 23.3% chance (100/4.3)3.1 is the equivalent of a 32.3% chance (100/3.1)Over-round = 102.1% (zero commission)

Calculating the over-round to lay2.2 is the equivalent of a 45.5% chance (100/2.2)4.4 is the equivalent of a 22.7% chance (100/4.4)3.25 is the equivalent of a 30.8% chance (100/3.25)Over-round = 99% (zero commission)

As with Betfair, it’s only the price about a Sweden win that shouldtempt backers away from the traditional bookmakers.

The 4.3 was commission-free with Betdaq, compared to acommission of up to 5% with Betfair on the 4.4. Not a bad offer,albeit to limited stakes, and certainly to be weighed closely againstthe SportsTab offer.

Finally, to the betting exchange located at GGBET.com.


England 2.1 2.11 2.2 2.25 2.25 2.30

(£5010) (£452) (£261) (£265) (£400) (£13)

Sweden 4.0 4.1 4.3 4.4 4.5 4.6

(£167) (£80) (£109) (£299) (£350) (£108)

Draw 2.9 3.0 3.1 3.2 3.3 3.4

(£87) (£1125) (£145) (£338) (£368) (£480)

Calculating the over-round to back2.2 is the equivalent of a 45.5% chance (100/2.2)4.3 is the equivalent of a 23.3% chance (100/4.3)3.1 is the equivalent of a 32.3% chance (100/3.1)Over-round = 101.1% (ignoring commission)

Calculating the over-round to lay2.25 is the equivalent of a 44.4% chance (100/2.25)4.4 is the equivalent of a 22.7% chance (100/4.4)3.2 is the equivalent of a 31.3% chance (100/3.2)

195


Over-round = 98.4% (ignoring commission)

Those seeking to lay an England win, a Sweden win or a draw inthe exchanges had the following best options to consider:

Lay England to win: At 2.25 with Betfair and/or GGBet, or at 2.2with Betdaq (zero commission on winnings).

Lay Sweden to win: At 4.5 with Betfair, or at 4.4 with GGBet and/or Betdaq (zero commission on winnings with the latter).

Lay the draw: At 3.2 with Betfair and/or GGBet or at 3.25 withBetdaq (zero commission on winnings with the latter).

Let’s see how this compares with what the professional layers(the bookmakers) were willing to offer.

Lay England to win: 2.25 (Paddy Power, Netbetsports); 2.3 (LionBets).

Exchange layers: 2.2 (Betdaq); 2.25 (Betfair, GGBet).In the best scenario, you could lay England to a total liability of

£265 with Betdaq at 2.2 (commission-free), and back them at 2.25 or2.3 with the bookmakers (tax-free). Even if you don’t want to hedgeyour bet this way, what the evidence shows is that you could havelaid a bet on an England win at shorter odds than were availablewith some established bookmakers.

Lay Sweden to win: 3.8 (Totalbet, Littlewoods); 4.35 (Lion Bets).Exchange layers: 4.4 (Betdaq, GGBet); 4.5 (Betfair).

Lay the draw: 3.2 (Littlewoods); 3.35 (SportsTab).Exchange layers: 3.2 (Betfair, GGBet); 3.25 (Betdaq).

Just for the record, I compiled the odds again with just 15 minutesto kick-off.

LadbrokesEngland to win: Evens (2.0)Sweden to win: 12 to 5 (3.4)Draw: 2 to 1 (3.0)

196

BETTING TO WIN

Over-round = 112.7% (unchanged)

Oddschecker best pricesEngland to win: 5 to 4 (2.25) – with Paddy Power, Sports.com,

Netbetsports.Sweden to win: 3 to 1 (4.0) – with LittlewoodsDraw: 21 to 10 (3.1) – with Littlewoods5 to 4 is the equivalent of a 44.4% chance. (100/2.25)3 to 1 is the equivalent of a 25% chance (100/4.0)21 to 10 is the equivalent of a 32.3% chance (100/3.1)

This adds up to a total implied chance for the three possibleoutcomes of 101.7%, down 0.3% from the previous evening.

What had happened was that the odds about a Sweden win atbest bookmaker prices had lengthened a little, and the odds about adraw had shortened a little. The net impact was to reduce the marginagainst the bettor from 2% to 1.7%.

BetbrainEngland to win: 2.3 – with Lion BetsSweden to win: 4.50 – with SportsTabDraw: 3.25 – with SportsTab, SportingbetUSA2.3 is the equivalent of a 43.5% chance (100/2.3)4.50 is the equivalent of a 22.2% chance (100/4.50)3.25 is the equivalent of a 30.8% chance (100/3.25)

This adds up to a total implied chance for the three possibleoutcomes of 96.5%, a tiny increase overnight.

On the Sunday morning, minutes before kick-off, the position onthe exchanges was as follows:

197


Betfair


England 2.1 2.15 2.2 2.25 2.3 2.35

(£3447) (£8426) (£289) (£7437) (£2394) (£850)

Sweden 4.1 4.2 4.3 4.4 4.5 4.6

(£27 (£1122) (£1117) (£160) (£766) (£1300)

Draw 2.95 3 3.1 3.2 3.3 3.4

(£33) (£652) (£952) (£3306) (£1065) (£500)

Since the prices posted the evening before, the only change (apartfrom the stake availabilities) was that layers were slightly better off(it was now possible to lay Sweden at 4.4 compared to 4.5 theprevious evening), but those seeking to back Sweden to win couldonly avail themselves of a price of 4.3 instead of the 4.4 previouslyavailable.

Betdaq


England 2.10 2.15 2.20 2.25 2.30 2.35

(£5452) (£1842) (£2432) (£1072) (£13) (£202)

Sweden 3.8 4.0 4.1 4.4 4.5 4.6

(£71) (£167) (£80) (£231) (£385) (£108)

Draw 2.9 3.0 3.1 3.2 3.25

(£87 (£1125) (£230) (£44) (£338)

Again, the odds to back Sweden to win had contracted since theprevious evening, although the odds available about an Englandwin had lengthened a little.

198

BETTING TO WIN

GGBet


England 2.1 2.11 2.2 2.25 2.3 2.4

(£200) (£195) (£27) (£379) (£100) (£200)

Sweden 3.7 3.9 4.1 4.3 4.4 4.6

(£310) (£66) (£74) (£104) (£89) (£100)

Draw 2.9 3.0 3.1 3.3 3.4 3.5

(£136) (£31) (£238) (£101) (£100) (£123)

Once again, the odds to back Sweden to win have contractedsince the previous evening, although the odds available about anEngland win were unchanged.

Taking all of these exchanges together, the situation with respectto backing and laying each of the options at best price, comparingthe immediate pre-match prices with those available the previousevening, was as follows:

Back England: Unchanged (2.2)Back Sweden: Better the previous evening (4.4 to 4.3)Back the draw: Unchanged (3.1)Lay England: Better the previous evening (2.2 to 2.25)Lay Sweden: Better pre-match (4.3 to 4.4)Lay the draw: Unchanged (3.2)

Calculating the over-round to back at best prices (ignoringcommission), immediately pre-match, we find:

2.2 is the equivalent of a 45.5% chance (100/2.2)4.3 is the equivalent of a 23.3% chance (100/4.3)3.1 is the equivalent of a 32.3% chance (100/3.1)Over-round = 101.1% (ignoring commission)

Calculating the over-round to lay at best prices (ignoringcommission), immediately pre-match the position was:

2.25 is the equivalent of a 44.4% chance (100/2.25)4.3 is the equivalent of a 23.3% chance (100/4.3)

199


3.2 is the equivalent of a 31.3% chance (100/3.2)Over-round = 99%

Calculating the over-round to back at best prices (ignoringcommission), the previous evening’s position showed:

2.2 is the equivalent of a 45.5% chance (100/2.2)4.4 is the equivalent of a 22.7% chance (100/4.4)3.1 is the equivalent of a 32.3% chance (100/3.1)Over-round = 100.5%

Calculating the over-round to lay at best prices (ignoringcommission), the previous evening, we get:

2.2 is the equivalent of a 45.5% chance (100/2.2)4.4 is the equivalent of a 22.7% chance (100/4.4)3.2 is the equivalent of a 31.3% chance (100/3.2)Over-round = 99.5% (ignoring commission)

In summary, ignoring stake availability, those seeking to backthe big game on the exchanges were, on the whole, better offdoing so the previous evening.

Those seeking to lay were better off doing so on the day,because of the slightly shorter price at which it was possible tolay Sweden.

Layers delay? That’s what the evidence of this match indicates,at least.

Spread BettingImmediately before the England v. Sweden game, Sporting Index,IG Index Sport, Cantor Sport, Spreadex and Sportsspread wereoffering the following quotes about the major markets:England supremacy:

0.2 – 0.5 (Sporting)0.3 – 0.6 (IG)0.4 (Cantor) (i.e. zero spread special offer)0.2 – 0.5 (Spreadex)0.3 – 0.6 (Sportsspread)

200

BETTING TO WIN

Total Goals:2.1 – 2.4 (Sporting)2.2 – 2.5 (IG)2.1 – 2.4 (Cantor)2.1 – 2.4 (Spreadex)2.4 – 2.7 (Sportsspread)

Shirt numbers:26 – 29 (Sporting)25 – 28 (IG)24 – 27 (Cantor)24 – 27 (Spreadex)24 – 27 (Sportsspread)

Bookings:36 – 40 (Sporting)36 – 40 (IG)34 – 38 (Cantor)33 – 37 (Spreadex)36 – 40 (Sportsspread)

Corners:10.25 – 11.25 (Sporting)10 – 11 (IG)9.5 – 10.5 (Cantor)9.5 – 10.5 (Spreadex)10 – 11 (Sportsspread)

First match goal:40 – 43 (Sporting)41 – 44 (IG)40 – 43 (Spreadex)37 – 40 (Sportsspread)

201


First England goal:52 – 55 (Sporting)51 – 54 (IG)51 – 54 (Spreadex)48 – 51 (Sportsspread)

First Sweden goal:60 – 63 (Sporting)61 – 64 (IG)58 – 61 (Spreadex)56 – 59 (Sportsspread)

25 – 10 – 0 Index (England)13.5 – 15 (Sporting)14 – 15.5 (IG)13 – 14.5 (Cantor)13 – 14.5 (Spreadex)13 – 14.5 (Sportsspread)

25 – 10 – 0 Index (Sweden)8.5 – 10 (Sporting)8.5 – 10 (IG)9 – 10.5 (Cantor)9 – 10.5 (Spreadex)9.5 – 11 (Sportsspread)

N.B. The 25 – 10 – 0 index (also known as the Windex) works asfollows:

25 points are awarded for a win10 points are awarded for a draw0 points are awarded for a loss

Let us examine each of these markets in turn, starting with thegoal supremacy market.

The first thing to look for is whether there exist what I have termed‘Quarbs’ (quasi-arbitrages). I divide Quarbs into two types: simpleQuarbs and full Quarbs.

202

BETTING TO WIN

Simple Quarbs exist when the top end of one company’s spread coincideswith the bottom end of another company’s spread. In the England v. Swedengame, no such situation existed in the supremacy market.

The second thing to look for is a full Quarb. This exists when oneof the spreads lies wholly outside the average of the mid-points ofall the spreads.

Applying this to the supremacy market produces the following:

England supremacy:

0.2 – 0.5 (Sporting) MID-POINT = 0.350.3 – 0.6 (IG) MID-POINT = 0.450.4 (Cantor) MID-POINT = 0.4 (Cantor was offering

a zero spread)0.2 – 0.5 (Spreadex) MID-POINT = 0.350.3 - 0.6 (Sportsspread) MID-POINT = 0.45

Average of the mid-points =(0.35 + 0.45 + 0.4 + 0.35 + 0.45)/5 =2/5= 0.4

0.4 lies within all the spreads, and is equal to the point spreadoffered by Cantor. There was, therefore, no possibility ofimplementing a trade based on Quarb strategy in this market.

The best option for those seeking to buy supremacy, however,was the 0.4 on offer with Cantor. Those seeking to sell again werebest off by selling at 0.4 with Cantor.

In the event, the sell proved the profitable trade. Was this thevalue bet before the match, and without the benefit of hindsight?This depends on what were objectively the real chances of anEngland win, a Sweden win and a draw.

One way of estimating this is to take the best odds on offer withthe betting exchanges, since the implied probabilities normally addup to about 100%, and reflects the combined wisdom of those willingto put their money where their mouth is. Let’s look at these again.

Calculating the over-round to back at best prices on the exchanges(as above):

England win: 1.2 to 1

203


Sweden win: 3.4 to 1Draw: 2.1 to 1Over-round = 101.1%

Do odds of 1.2 to 1 about an England win translate into an Englandsupremacy rating of less than 0.4 goals? Clearly England, as favourites,start with a positive goal supremacy rating, but whether value existsin SELL at 0.4 in these circumstances really is a close call.

Let us turn now to the total goals market.

Total goals:2.1 – 2.4 (Sporting)2.2 – 2.5 (IG)2.1 – 2.4 (Cantor)2.2 – 2.5 (Spreadex)2.4 – 2.7 (Sportsspread)

Again, the first thing to look for is whether there exist what Ihave termed ‘Quarbs’ (quasi-arbitrages).

You may remember that simple Quarbs exist when the top end ofone company’s spread coincides with the bottom end of anothercompany’s spread. In the England v. Sweden game, this existedbetween Sporting and Cantor (2.1 – 2.4) and Sportsspread (2.4 –2.7). This means that you could buy with one company and sellwith another, and make a certain profit (loss) of zero.

This is indicative, at least, that one of the companies is a little outin its judgement, and it may be possible to take advantage. To see ifthis is the case, we need to determine whether a full Quarb exists.This exists when one of the spreads lies wholly outside the averageof the mid-points of all the spreads.

Applying this analysis to the total goals market produces thefollowing:

Total Goals2.1 – 2.4 (Sporting) MID-POINT = 2.252.2 – 2.5 (IG) MID-POINT = 2.352.1 – 2.4 (Cantor) MID-POINT = 2.252.2 – 2.5 (Spreadex) MID-POINT = 2.35

204

BETTING TO WIN

2.4 – 2.7 (Sportsspread) MID-POINT = 2.55Average of the mid-points = (2.25 + 2.35 + 2.25 + 2.35 + 2.55)/5

= 11.75/5 = 2.35Sportsspread’s quote of 2.4 – 2.7 lies everywhere outside the

average mid-point (2.35).In other words, we are saying that the market-makers of the five

companies have produced a combined best estimate that theexpected number of goals in the match will be 2.35.

If this is truly the best estimate, then the value exists in a sell oftotal goals at 2.4 with Sportsspread. In the event, a sell at 2.4 didprove a profitable trade.

Turning now to the shirts market, which quotes an aggregate ofthe shirt numbers worn by players who score.

Shirt numbers:26 – 29 (Sporting)25 – 28 (IG)24 – 27 (Cantor)24 – 27 (Spreadex)24 – 27 (Sportsspread)The average of the mid-points is:

(27.5 + 26.5 + 25.5 + 25.5 + 25.5)/5 = 26.1There is almost a full Quarb with Sporting, but not quite.

Bookings:36 – 40 (Sporting)36 – 40 (IG)34 – 38 (Cantor)33 – 37 (Spreadex)36 – 40 (Sportsspread)The average of the mid-points is: (38 + 38 + 36 + 35 + 38)/5 = 37As near as it is possible to get to a full Quarb (courtesy of Spreadex),

but not quite.

205


Corners:10.25 – 11.25 (Sporting)10 – 11 (IG)9.5 – 10.5 (Cantor)9.5 – 10.5 (Spreadex)10 – 11 (Sportsspread)The average of the mid-points is:

(10.75 + 10.5 + 10 + 10 + 10.5)/5 = 51.75/5= 10.35

No Quarb.

Time to first match goal (minutes):40 – 43 (Sporting)41 – 44 (IG)40 – 43 (Spreadex)37 – 40 (Sportsspread)

There is clearly a simple Quarb here, in that the top end of theSportsspread quote (40) coincides with the bottom end of the Sportingquote (40). This means that you could buy with Sportsspread and sellwith Sporting, and make a certain profit (loss) of zero.

Again, this is indicative that one of the companies may be a littleout in its judgement. To see if this is the case, we need to determinewhether a full Quarb exists.

Applying this analysis to the first match goal market producesthe following:

Time to first match goal40 – 43 (Sporting) MID-POINT = 41.541 – 44 (IG) MID-POINT =42.540 – 43 (Spreadex) MID-POINT = 41.537 – 40 (Sportsspread) MID-POINT = 38.5Average of the mid-points = (41.5 + 42.5 + 41.5 + 38.5)/4 = 41Sportsspread’s quote of 37 – 40 lies everywhere outside the average

mid-point (41).In other words, we are saying that the market-makers of these

companies have produced a combined best estimate that the time

206

BETTING TO WIN

of the first match goal will be 41 minutes. If this is truly the bestestimate, then the value exists in a buy of minutes to the first matchgoal at 40, with Sportsspread.

The most sophisticated bettors, however, will be aware of whatwas discussed in an earlier chapter, i.e. that the time of first goalmarkets are often set artificially high, to accommodate thepreponderance of buyers in the market. Allowing for this, any tradewould offer marginal value at best.

Time to first England goal (minutes):52 – 55 (Sporting)51 – 54 (IG)51 – 54 (Spreadex)48 – 51 (Sportsspread)

This is the interesting case of an outright arbitrage. An arbitrage(or ‘arb’) exists when it is possible to buy with one company and tosell with another, and make a profit regardless of the outcome. Inthis example, you could buy minutes at 51 with Sportsspread, andsell minutes at 52 with Sporting, and win whatever the outcome.

As an example, if England scored after one minute, you wouldlose 50 times your stake with Sportsspread, and win 51 times yourstake with Sporting. At the other end of the scale, if England scoredafter 90 minutes (or didn’t score at all), the market makes up at 90,and you would win 39 times your stake with Sportsspread (lose 38times your stake with Sporting).

The problems with arbs are outlined in an earlier chapter. Basically,repeated trading at these prices will either make you very unpopularwith the companies concerned, with associated consequences, orelse your stake is liable to be severely curtailed if you are allowed totrade at the arb position at all. This market is probably best avoided,at least until and unless the arb turns into a Quarb. It is worth notingin this context that spread betting companies tend to assume thatthose who systematically trade when ‘arbs’ exist are taking bothsides of the ‘arb’, whether they are or not.

First Sweden goal:60 – 63 (Sporting)

207


61 – 64 (IG)58 – 61 (Spreadex)56 – 59 (Sportsspread)

Again we have a straight arb between Sporting and Sportsspread.For the reasons mentioned above this is probably best avoided again.

25 – 10 – 0 Index (England)13.5 – 15 (Sporting)14 – 15.5 (IG)13 – 14.5 (Cantor)13 – 14.5 (Spreadex)13 – 14.5 (Sportsspread)The mid-points average is:

(14.25 + 14.75 + 13.75 + 13.75+ 13.75)/5 = 14.05Almost a full Quarb, given a sell of England at 14 with IG, but not

quite.

25 – 10 – 0 Index (Sweden)8.5 – 10 (Sporting)8.5 – 10 (IG)9 – 10.5 (Cantor)9 – 10.5 (Spreadex)9.5 – 11 (Sportsspread)The average of the mid-points is:

(9.25 + 9.25 + 9.75 + 9.75 + 10.25)/5 = 9.65No Quarb.

Exchange Spread BettingAnother avenue open to those attracted by spread betting is that ofthe betting exchanges. Let’s examine some of the options.

The Goal Supremacy MarketIn the minutes before the kick-off in the England v. Sweden game,it was possible to sidestep the spread firms and to trade the rangemarkets instead on Betfair. The quotes were as follows:

208

BETTING TO WIN

Sell England supremacy @ 0.4 for up to £65.Buy England supremacy @ 0.45 for up to £5; at 0.5 up to £324; at

0.7 up to £50.The market was bounded between - 3 goals and + 4 goals, in

order to set a maximum limit on potential liabilities. The tradingunit was 0.05 of a goal.

Since it was possible to sell England supremacy at 0.4 with Cantor,commission-free, the only possible advantage to playing this marketin the exchanges was the limited liability imposed by the operators.If you are really worried about a 12 – 0 scoreline, you could limityour stakes instead, although that would also limit your potentialprofit.

The Bookings MarketOn Betfair, the quotes were as follows:

Sell bookings @ 31 for up to £1.Buy bookings @ 36 for up to £10; at 38 up to £8; at 39 up to £20.

The market was bounded between 0 and 120 bookings points.The trading unit was 1 point.

There is clearly no advantage at all in a sell at 31, since you couldsell with Sporting, IG and Sportsspread at 36, commission-free.

You could buy at 36, instead of a best price of 37, available withSpreadex. However, you would be limited to £10 a point, you wouldpay up to 5% commission on any winnings, and you would besubject to a maximum ceiling of 120 bookings points however manycards were issued. The best strategy is apparently to buy at 36 withBetfair, or at 37 with Spreadex.

To see the relative merits of each option, consider just the followingpossible outcomes, and assume a unit stake of £10:

a. No bookings pointsb. 20 bookings pointsc. 40 bookings pointsd. 100 bookings pointse. 200 bookings points

209


No bookings pointsLoss of £360 (Betfair)Loss of £370 (Spreadex)

20 bookings pointsLoss of £160 (Betfair)Loss of £170 (Spreadex)

40 bookings pointsProfit of £40 on Betfair (£38 after deducting 5% commission)Profit of £30 on Spreadex



Clearly, therefore, the higher the number of bookings you expect,the more favourably you would consider the bet at 37 on Spreadex,and notably if you expect it to top 120 bookings points.

The Corners MarketOn Betfair the quotes were as follows:

Sell corners @ 10.5 for up to £12; at 10.4 up to £22; at 9.5 up to£37.

Buy corners @ at 10.6 for up to £35; at 10.7 up to £14; at 10.8 up to£203.

The market was bounded between 7 and 17 corners. The tradingunit was 0.1 corners.

There is clearly no advantage in a buy of corners at 10.6 withBetfair when you could buy corners, commission-free, at 10.5with Cantor or Spreadex, unless you want to be protected

210

BETTING TO WIN

against the possibility of a corners total of less than seven. Analternative strategy, if you are averse to risk, is to reduce yourunit stake.

The Total Goals MarketOn Betfair the quotes were as follows:

Sell goals @ 2.25 for up to £172; at 2.05 up to £2.Buy goals @ 2.3 for up to £305; at 2.4 up to £15; at 2.45 up to £66.

The market was bounded between 0 and 6 goals. The tradingunit was 0.05 goals.

There was clearly no advantage to a sell of total goals at 2.25, sinceyou could sell at 2.4 with Sportsspread, unless you wanted to lockin the maximum liability provision of the six-goal maximum onBetfair.

The best buy price among the spread companies was the 2.4, withSporting and Cantor, which compared unfavourably with thepossibility of a buy at 2.3 with Betfair.

However, you would pay up to 5% commission on any winnings,and you would be subject to a maximum ceiling of six goals.

Therefore we need to consider whether to buy at 2.3 with Betfair,or at 2.4 with Sporting and Cantor.

To see the relative merits of each option, consider just the followingpossible outcomes, and assume a unit stake of £100 a goal:

a. No goalsb. 3 goalsc. 6 goalsd. 7 goals

No goalsLoss of £230 on BetfairLoss of £240 on Spreadex

3 goalsProfit of £70 on Betfair (£66.50 after deducting 5% commission)Profit of £60 on Spreadex

211


6 goalsProfit of £370 on Betfair (£351.50 after deducting 5% commission)Profit of £360 on Spreadex

7 goalsProfit of £370 on Betfair (£351 - 50 after deducting 5% commission)Profit of £460 on Spreadex

Clearly, therefore, the greater the number of goals you expect,the more favourably you would consider the bet at 2.4 on Spreadex,and notably if you expect the goal total to top six.

212

BETTING TO WIN

Chapter 2: If It Looks Too Good To Be

True, Then It Is Too Good To Be True!

The heading of the final Chapter is not meant to discourage youfrom spotting genuine value, but to caution against traps whichcan easily ensnare the unwary.

A classic example is the first goal line market offered by Betfair.This is a line market, and not a range market. In other words,you are invited to bet whether the first goal will be scored beforeor after a given time. You win the equivalent of your stake (andyour stake returned) if you guess correctly. You lose your stake ifyou guess wrongly. This is very different from a spread marketwhere you are invited to buy or sell at a given level, and whereyour winnings/losses are determined by the extent to which youare right or wrong.

The key difference in these two types of bet is that one requiresan estimate of the median and the other an estimate of the mean.

The median is the average in the sense that there is an equalchance of the outcome being greater than or less than it. Take,for example, the following fictional list of first goal times in a setof nine matches.

Match 1 – 30 minutesMatch 2 – 8 minutesMatch 3 – 84 minutesMatch 4 – 52 minutesMatch 5 – 70 minutesMatch 6 – 32 minutesMatch 7 – 18 minutesMatch 8 – 10 minutesMatch 9 – 29 minutes

The median time of the first goal is 30 minutes, since there arean equal number of goals scored before and after this time.

8, 10, 18 and 29 are less than 30; 32, 52, 70 and 84 are greaterthan 30.

213

IF IT LOOKS TOO GOOD TO BE TRUE, THEN IT IS TOO GOOD TO BE TRUE!

If there are an even number of observations, there are two middlepoints. To calculate the median, add the two together and divide bytwo.

The mean time of the first goal is, however, the average time ofall the first goals, in the sense of all the times added up anddivided by the number of goals.

In this example, the mean time of the first goal is30 + 8 + 84 + 52 + 70 + 32 + 18 + 10 + 29 = 333/9=37 minutes.In other words, the median is less than the mean in this

example. For any goal likely to be scored in the first half of playingtime, this is likely to be true, as the mean is skewed by a few veryhigh numbers. The converse applies to goals likely to be scoredin the second half of playing time.

For the purpose of spread betting, it is clear that the mean isthe important number, as a small number of very low or highnumbers are important in determining your final profit or loss.

For the purpose of line betting, however, it is the median whichis important, as a very high number is no more significant than amuch lower number, as long as they are both above or both belowthe level at which you traded.

It is important not to confuse these two numbers in formulatingyour betting strategy in line and range (or spread) betting marketsrespectively. To illustrate the issue, I have collated the times of allthe goals scored in World Cup 2002 (see Appendix 2).

We would expect, as explained above, that for a series of goalsscored in the first half of play (such as the first goal in each match)that the median would tend to be less than the mean.

In fact, the relevant calculation produces the following results:Median time of 1st goal = 32 minutes; mean time of 1st goal =

36 minutes.For comparison purposes, the median time of the first goals

scored in the Premiership in the 2001 – 2002 season was 28minutes, whereas the mean time of the first goal scored was 35.4minutes.

In conclusion, it is important to distinguish between mean times(suitable for trading in the ‘time of first goal’ spread or rangemarkets) and median times (suitable for trading in the ‘time offirst goal’ line markets).

214

BETTING TO WIN

Trade accordingly.I have also collated (Appendix 2) the times of all the disciplinary

cards issued in World Cup 2002. For the particularly assiduousreader, a useful exercise would be to use these data similarly tocalculate the difference between mean and median times in thebookings market.

Finally, an interesting and potentially useful perspective of goaltimes can be gleaned courtesy of what is known in statisticalcircles as the Poisson distribution. Originally, it was formulated tomodel the rate at which Prussian officers were kicked to deathby horses. Nowadays it has a more prosaic use, i.e. to calculatethe likelihood of a given number of goals being scored by eachteam. It works like this.

First, we need an estimate of the number of goals that a teamwill score in a given match. From this estimate, the Poissondistribution allows us to estimate the chance of any game finishingwith any particular scoreline.

Get this right, and there are potentially rich pickings availableacross a number of markets.

The Poisson distribution is included in a table (see Appendix3) which estimates the percentage chance of a team scoring agiven number of goals based on our estimate of the number theyare likely to score. The value of the table lies in the fact that thepredictions of the Poisson distribution tend to conform to actualresults over time.

For example, if we estimate that Southampton will, on average,score one goal against Arsenal, then the Poisson distribution tells usthat there is a 36.8% chance that they will not score any goals onthis particular occasion. It tells us that there is a 73.6% chance of onegoal or less, a 92.0% chance of 2 goals or less, a 98.1% chance of 3goals or less, and so on. This can be read off directly from the table.Look down the vertical column on the left of the table to 1.0, andread across the horizontal row to find the probability of any givennumber of goals (or less) being scored. From this table we can evenwork out the probability of exactly 3 goals being scored, say. To dothis, we read off the likelihood of 2 goals or less being scored (92.0%),and then the likelihood of 3 goals or less being scored (98.1%). Thedifference between these numbers (6.1%) is the likelihood of exactly

215

IF IT LOOKS TOO GOOD TO BE TRUE, THEN IT IS TOO GOOD TO BE TRUE!

3 goals being scored.Similarly, the chance of Southampton scoring exactly one goal

is equal to the chance of them scoring 1 goal or fewer (73.6%)and the chance of them scoring no goals (36.8%). This works outat 36.8%.

To take another example, we estimate that Arsenal will score,on average, 1.5 goals against Southampton. We can now use thePoisson table to estimate the likelihood of Arsenal scoring no goalson this occasion, or one, or indeed any number of your choice.

As before, read down to 1.5. Then read across. This tells us thefollowing:

Arsenal chance of scoring no goals = 22.3%Arsenal chance of scoring 1 goal or less = 55.8%Arsenal chance of scoring 2 goals or less = 80.9%Arsenal chance of scoring 3 goals or less = 93.4%Arsenal chance of scoring 4 goals or less = 98.1%Arsenal chance of scoring 5 goals or less = 99.6%Arsenal chance of scoring 6 goals or less = 99.9%

We can use the Poisson table to estimate the true probability ofany given scoreline, although to do so we must make the notcompletely realistic assumption that the number of goals scoredby one team is independent of the number of goals scored by theother. Not completely realistic, but the results of adopting thisassumption seem to work out reasonably well in practice.

On this basis, if we want to estimate the likelihood of a 2 – 1scoreline to Arsenal, it’s as simple as this:

First, calculate the probability of Arsenal scoring exactly 2 goals.From Appendix 3, the chance of Arsenal scoring 2 goals or less

is 80.9%, and the chance of Arsenal scoring 1 goal or less is 55.8%.Therefore, it follows that the chance that Arsenal will score exactly2 goals is the difference between these figures, i.e. 25.1%.

Second, calculate the probability of Southampton scoring 1 goalexactly. This is 36.8% (same reasoning as above).

Now multiply these together.25.1% × 36.8% = 9.24%.Thus the chance of a 2 – 1 scoreline to Arsenal, according to

216

BETTING TO WIN

the workings of the Poisson table, is 9.24%.This can be converted into odds quite simply: (100 - 9.24)/9.24

= 90.76/9.24 = 9.82 to 1Now that you know how to estimate the true odds, compare

these with the abundance of odds actually on offer in the variousforums. Another arrow in the quiver.

217

DRAW BIASES

Conclusion

In this book, I have drawn upon the experience I have built up overmany years of professional study of betting markets.

In particular, I have investigated and sought to make accessible toa wider audience the results of an extensive body of workcontributed by experts around the world, in order to study the wayin which bettors really can turn the odds in their favour. Much ofthis work has, until known, been relatively unknown outside ofacademic circles.

In doing so, I have sought to distinguish what is valid and usefulfrom what is merely speculative or just plain wrong.

I have tried to do this in an entertaining and topical, as well as aninformative, manner. In this quest, I hope I have succeeded.

The Golden Age of betting is now. We are living in it. I hope thatthis book will help you to enjoy and profit from our good fortune tothe full.

218

BETTING TO WIN

Appendix 1: Some Interesting And

Useful Internet Sites

These are not arranged in any special order and are all included ingood faith. As a general rule, however, bettors should always beware.Internet bookmakers do go out of business still owing money toclients.

Be careful in judging whether to bet with a firm.

Bookmakerswww.tote.co.uk

www.bluesq.com

www.paddypower.com

www.sportingbet.com

www.eurobet.com

www.gamebookers.com

www.ukbetting.com

www.willhill.com

www.sportingodds.com

www.ladbrokes.com

www.skybet.com

www.bet365.co.uk

www.betinternet.com

www.intertops.com

www.sportsinteraction.com

www.sportstab.com.au

www.centrebet.com

www.expekt.com

www.betandwin.com

www.stanjames.com

www.canbet.com.au

www.victorchandler.com

www.thegreek.com

www.planetpinnacle.com

www.betabet.com

www.yabet.com

219

INTERESTING AND USEFUL INTERNET SITES

www.totalbet.com

www.xodds.com

www.attheraces.co.uk

www.betonmarkets.com

www.betdirect.net

Spread Betting Companieswww.sportingindex.com

www.cantorindex.com

www.igindex.co.uk

www.igsport.com

www.cantorsport.com

www.spreadex.com

www.spreadexfinancials.com

www.cityindex.co.uk

www.finspreads.com

www.tradindex.com

www.deal4free.com

Betting Exchangeswww.betfair.com

www.betdaq.com

www.betdaq.co.uk

www.ggbet.com (see also www.iwageru.co.uk)

www.sportingoptions.co.uk

www.tradindex.com (sports)

www.intrade.com

www.newsfutures.com

www.biz.uiowa.edu/iem

www.tradesports.com

www.tradingsports.com

Sports and Betting Information and Advicewww.highstakes.co.uk

www.racingpost.co.uk

www.sportinglife.com

www.skysports.com

220

BETTING TO WIN

www.bbc.co.uk/sport

www.football365.com

www.readabet.com

www.oddschecker.com

www.oddschecker.co.uk

www.betbrain.com

www.easyodds.com

www.bestbetting.com

www.sportsbettingindex.com

www.soccerbase.com

www.sportal.com

www.smartbet.co.uk

www.attheraces.co.uk

www.gamebookers.com

www.spreadtrades.com

www.bookiesindex.com

www.bettingadvice.com

www.crastinum.com

www.bookiebusters.net

www.soccernet.com

www.onewaybet.com

www.opta.co.uk

www.bethelp.com

221

WORLD CUP 2002 STATISTICS

Appendix 2: World Cup 2002

Statistics

Times of goals (+ indicates stoppage time, where relevant)Times of cards (+ indicates stoppage time, where relevant)

For betting analysis purposes, we treat 45 as the maximum for afirst half goal/card, and 90 as the maximum for a second half goal/card. Goals/cards after full time are shown but excluded for bettinganalysis purposes.

When a red card results from a second yellow, only the time ofthe red card is shown.

N.B. Where no goals are scored in a match, this is normally classedas 90 for ‘time of first goal’ purposes.

Group A

FRANCE v. SENEGALScore: 0 – 1Goal times: 30Yellow card times: 47+ (i.e. 45), 51Red card times: –

URUGUAY v. DENMARKScore: 1 – 2Goal times: 45, 47, 83Yellow card times: 25, 34, 51Red card times: –

FRANCE v. URUGUAYScore: 0 – 0Goal times: –Yellow card times: 11, 47+ (i.e. 45), 47+ (i.e. 45), 48+ (i.e. 45), 47Red card times: 25

222

BETTING TO WIN

DENMARK v. SENEGALScore: 1 – 1Goal times: 16, 52Yellow card times: 7, 10, 20, 62, 82, 84Red card times: 80

DENMARK v. FRANCEScore: 2 – 0Goal times: 22, 67Yellow card times: 27, 71, 81Red card times: –

SENEGAL v. URUGUAYScore: 3 – 3Goal times: 20, 26, 38, 46, 69, 88Yellow card times: 2, 4, 8, 19, 35, 39, 40, 69, 82, 82, 87, 87Red card times: –

Group B

PARAGUAY v. SOUTH AFRICAScore: 2 – 2Goal times: 39, 55, 63, 91(i.e. 90)Yellow card times: 3, 9, 35, 38, 47+ (i.e. 45), 65, 90, 93 (i.e. 90)Red card times: –

SPAIN v. SLOVENIAScore: 3 – 1Goal times: 44, 74, 82, 87Yellow card times: 36, 46+ (i.e. 45), 65Red card times: –

SPAIN v. PARAGUAYScore: 3 – 1Goal times: 10, 53, 69, 83Yellow card times: 9, 44, 60, 80Red card times: –

223


SOUTH AFRICA v. SLOVENIAScore: 1 – 0Goal times: 4Yellow card times: 12, 35, 52, 62, 75Red card times: –

SOUTH AFRICA v. SPAINScore: 2 – 3Goal times: 4, 31, 46+ (i.e. 45), 53, 56Yellow card times: 16, 35, 52, 62, 67, 69, 75, 81Red card times: –

SLOVENIA v. PARAGUAYScore: 1 – 3Goal times: 46+ (i.e. 45), 65, 73, 84Yellow card times: 4, 15, 68, 69, 79Red card times: 22 (2nd yellow), 81

Group C

BRAZIL v. TURKEYScore: 2 – 1Goal times: 47+ (i.e. 45), 50, 87Yellow card times: 21, 24, 44, 73,Red card times: 86, 94 (i.e. 90) (2nd yellow)

CHINA v. COSTA RICAScore: 0 – 2Goal times: 61, 65Yellow card times: 15, 17, 60, 72, 77, 79, 85Red card times: –

BRAZIL v. CHINAScore: 4 – 0Goal times: 15, 32, 45, 55Yellow card times: 25, 69Red card times: –

224

BETTING TO WIN

COSTA RICA v. TURKEYScore: 1 – 1Goal times: 56, 86Yellow card times: 20, 24, 43, 45, 89Red card times: –

COSTA RICA v. BRAZILScore: 2 – 5Goal times: 10, 13, 38, 39, 56, 62, 64Yellow card times: 93 (i.e. 90)Red card times: –

TURKEY v. CHINAScore: 3 – 0Goal times: 6, 9, 85Yellow card times: 19, 30, 46+ (i.e. 45), 62, 81Red card times: 58

Group D

SOUTH KOREA v POLANDScore: 2 – 0Goal times: 26, 53Yellow card times: 31, 70, 79, 84, 90Red card times: –

USA v. PORTUGALScore: 3 – 2Goal times: 4, 29, 36, 39, 71Yellow card times: 34, 52, 92 (i.e. 90)Red card times: –

SOUTH KOREA v. USAScore: 1 – 1Goal times: 24, 78Yellow card times: 30, 39, 80Red card times: –

225


PORTUGAL v. POLANDScore: 4 – 0Goal times: 14, 65, 77, 88Yellow card times: 21, 25, 27, 31, 39Red card times: –

PORTUGAL v. SOUTH KOREAScore: 0 – 1Goal times: 70Yellow card times: 22, 24, 57, 74, 83, 93 (i.e. 90)Red card times: 27, 66 (2nd yellow)

POLAND v. USAScore: 3 – 1Goal times: 3, 5, 66, 83Yellow card times: 44, 46, 63, 72, 86Red card times: –

Group E

IRELAND v. CAMEROONScore: 1 – 1Goal times: 39, 52Yellow card times: 30, 51, 82, 89Red card times: –

GERMANY v. SAUDI ARABIAScore: 8 – 0Goal times: 20, 25, 40, 46+ (i.e. 45), 70, 73, 84, 91 (i.e. 90)Yellow card times: 43, 83, 91 (i.e. 90)Red card times: –

GERMANY v. IRELANDScore: 1 – 1Goal times: 19, 92 (i.e. 90)Yellow card times: –Red card times: –

226

BETTING TO WIN

CAMEROON v. SAUDI ARABIAScore: 1 – 0Goal times: 66Yellow card times: 10, 59Red card times: –

CAMEROON v. GERMANYScore: 0 – 2Goal times: 50, 79Yellow card times: 8, 9, 29, 31, 37, 42, 42, 44, 56, 58, 60, 72, 74, 81Red card times: 40 (2nd yellow), 77 (2nd yellow)

SAUDI ARABIA v. IRELANDScore: 0 – 3Goal times: 7, 61, 87Yellow card times: 61, 70Red card times: –

Group F

ENGLAND v. SWEDENScore: 1 – 1Goal times: 24, 59Yellow card times: 12, 47+ (i.e. 45), 73Red card times: –

ARGENTINA v. NIGERIAScore: 1 – 0Goal times: 63Yellow card times: 51, 73, 90Red card times: –

SWEDEN v. NIGERIAScore: 2 – 1Goal times: 27, 35, 63Yellow card times: 31, 69, 80Red card times: –

227


ARGENTINA v. ENGLANDScore: 0 – 1Goal times: 44Yellow card times: 13, 29, 50Red card times: –

SWEDEN v. ARGENTINAScore: 1 – 1Goal times: 59, 88Yellow card times: 55, 58, 65, 75, 78Red card times: 47+ (i.e. 45)

NIGERIA v. ENGLANDScore: 0 – 0Goal times: –Yellow card times: –Red card times: –

Group G

CROATIA v. MEXICOScore: 0 – 1Goal times: 60Yellow card times: –Red card times: 60

ITALY v. ECUADORScore: 2 – 0Goal times: 7, 27Yellow card times: 14, 49, 54Red card times: –

ITALY v. CROATIAScore: 1 – 2Goal times: 55, 73, 76Yellow card times: 39, 51Red card times: –

228

BETTING TO WIN

MEXICO v. ECUADORScore: 2 – 1Goal times: 5, 28, 57Yellow card times: 15, 27, 49, 61, 65, 87Red card times: –

MEXICO v. ITALYScore: 1 – 1Goal times: 34, 85Yellow card times: 2, 5, 10, 43, 55, 57, 84Red card times: –

ECUADOR v. CROATIAScore: 1 – 0Goal times: 48Yellow card times: 72, 86, 92 (i.e. 90)Red card times: –

Group H

JAPAN v. BELGIUMScore: 2 – 2Goal times: 57, 59, 67, 75Yellow card times: 21, 31, 54, 62, 82Red card times: –

RUSSIA v. TUNISIAScore: 2 – 0Goal times: 59, 64Yellow card times: 27, 50, 75, 88Red card times: –

JAPAN v. RUSSIAScore: 1 – 0Goal times: 51Yellow card times: 13, 15, 38, 42, 60, 91 (i.e. 90)Red card times: –

229


TUNISIA v. BELGIUMScore: 1 – 1Goal times: 13, 17Yellow card times: 22, 40, 43, 68, 69Red card times: –

TUNISIA v. JAPANScore: 0 – 2Goal times: 48, 75Yellow card times: 21, 81Red card times: –

BELGIUM v. RUSSIAScore: 3 – 2Goal times: 7, 52, 78, 82, 88Yellow card times: 12, 14, 39, 64, 84Red card times: –

Round of 16

GERMANY v. PARAGUAYScore: 1 – 0Goal times: 88Yellow card times: 26, 35, 50, 71, 92 (i.e. 90)Red card times: 92 (i.e. 90)

DENMARK v. ENGLANDScore: 0 – 3Goal times: 5, 22, 44Yellow card times: 24, 50Red card times: –

SWEDEN v. SENEGALScore: 1 – 1 (1 – 2 after extra time)Goal times: 11, 37, 104 (excluded)Yellow card times: 73, 94 (excluded)Red card times: –

230

BETTING TO WIN

SPAIN v. IRELANDScore: 1 – 1 (Match eventually decided on penalties)Goal times: 8, 90Yellow card times: 62, 87, 89Red card times: –

MEXICO v. USAScore: 0 – 2Goal times: 8, 65Yellow card times: 26, 37, 47, 50, 53, 67, 70, 81, 83, 84Red card times: 88

BRAZIL v. BELGIUMScore: 2 – 0Goal times: 67, 87Yellow card times: 24, 28Red card times: –

JAPAN v. TURKEYScore: 0 – 1Goal times: 12Yellow card times: 21, 44, 45, 91(i.e. 90)Red card times: –

S. KOREA v. ITALYScore: 1 – 1 (2 – 1 after extra time)Goal times: 18, 88, 117 (excluded)Yellow card times: 4, 17, 22, 55, 59, 80, 99 (excluded), 115 (excluded)Red card times: 103 (2nd yellow) (excluded)

231


Quarter-finals

ENGLAND v BRAZILScore: 1 – 2Goal times: 23, 47+ (i.e. 45), 50Yellow card times: 75, 86Red card times: 57

GERMANY v. USAScore: 1 – 0Goal times: 39Yellow card times: 40, 41, 66, 68, 68, 69, 70Red card times: –

SPAIN v. SOUTH KOREAScore: 0 – 0 (3:5 after penalty shoot-out)Goal times: –Yellow card times: 52, 53, 111 (excluded)Red card times: –

SENEGAL v. TURKEYScore: 0 – 1Goal times: 94 (i.e. 90)Yellow card times: 12, 22, 63, 87Red card times: –

Semi-finals

GERMANY v SOUTH KOREAScore: 1 – 0Goal times: 75Yellow card times: 71, 85, 94 (i.e. 90)Red card times: –

232

BETTING TO WIN

BRAZIL v. TURKEYScore: 1 – 0Goal times: 49Yellow card times: 41, 59, 90Red card times: –

3rd place play-off

SOUTH KOREA v. TURKEYScore: 2 – 3Goal times: 1, 9, 13, 32, 93Yellow card times: 23, 50, 83Red card times: –

World Cup final

GERMANY v. BRAZILScore: 0 – 2Goal times: 67, 79Yellow card times: 6, 9Red card times: –

233


Appendix 3:Poisson Table

chance of scoring x goals or less; x =

Expected 0 1 2 3 4 5 6

No. of goals

0.1 0.905 0.995 1.000 1.000 1.000 1.000 1.000

0.2 0.819 0.982 0.999 1.000 1.000 1.000 1.000

0.3 0.741 0.963 0.996 1.000 1.000 1.000 1.000

0.4 0.670 0.938 0.992 0.999 1.000 1.000 1.000

0.5 0.607 0.910 0.986 0.998 1.000 1.000 1.000

0.6 0.549 0.878 0.977 0.997 1.000 1.000 1.000

0.7 0.497 0.844 0.966 0.994 0.999 1.000 1.000

0.8 0.449 0.809 0.953 0.991 0.999 1.000 1.000

0.9 0.407 0.772 0.937 0.987 0.998 1.000 1.000

1.0 0.368 0.736 0.920 0.981 0.996 0.999 1.000

1.1 0.333 0.699 0.900 0.974 0.995 0.999 0.999

1.2 0.301 0.663 0.880 0.966 0.992 0.999 1.000

1.3 0.273 0.627 0.857 0.957 0.989 0.998 1.000

1.4 0.247 0.592 0.834 0.946 0.986 0.997 0.999

1.5 0.223 0.558 0.809 0.934 0.981 0.996 0.999

1.6 0.202 0.525 0.783 0.921 0.976 0.994 0.999

1.7 0.183 0.493 0.757 0.907 0.970 0.992 0.998

1.8 0.165 0.463 0.731 0.891 0.964 0.990 0.997

1.9 0.150 0.434 0.704 0.875 0.956 0.987 0.997

2.0 0.135 0.406 0.677 0.857 0.947 0.983 0.996

2.2 0.111 0.355 0.626 0.819 0.928 0.975 0.993

2.4 0.091 0.308 0.570 0.779 0.904 0.964 0.988

2.6 0.074 0.267 0.518 0.736 0.877 0.951 0.983

2.8 0.061 0.231 0.470 0.692 0.848 0.935 0.976

3.0 0.050 0.199 0.423 0.647 0.815 0.916 0.967

3.2 0.041 0.171 0.380 0.603 0.781 0.895 0.955

3.4 0.033 0.147 0.340 0.558 0.744 0.871 0.942

3.6 0.027 0.126 0.303 0.515 0.706 0.844 0.927

234

BETTING TO WIN

Bibliography

Ali, M.M. (1977) Probability and utility estimates for racetrack bettors, Journalof Political Economy, 83, 803 – 815.

Ali, M.M. (1979) Some evidence on the efficiency of a speculative market,Econometrica, 47, 387 – 392.

Asch, P. and Quandt, R.E. (1986) The Professor’s Guide to Strategies. AuburnHouse, Dover, MA.

Asch, P. and Quandt, R.E. (1987) Efficiency and profitability in exotic bets,Economica, 59, 278 – 298.

Asch, P. and Quandt, R.E. (1988) Betting bias in exotic bets, Economic Letters,28, 215 – 219.

Avery, C. and Chevalier, J. (1999), ‘Identifying investor sentiment from pricepaths: the case of football betting, Journal of Business, 72 (4), 493 –521.

Benter, W. (1994) Computer based horse race handicapping and wageringsystems: a report. In: Efficiency of Racetrack Betting Markets (1994),eds Hausch, D.B., Lo, S.Y. and Ziemba, W.T., pp. 183 – 198, AcademicPress, London.

Betton, S. (1994) Post position bias: an econometric analysis of the 1987season at Exhibition Park. In: Efficiency of Racetrack Betting Markets(1994), eds Hausch, D.B., Lo, S.Y. and Ziemba, W.T., pp. 511 – 526,Academic Press, London.

Bird, R. and McCrae, M. (1987) Tests of the efficiency of racetrackbetting using bookmakers’ odds, Management Science, 33, 1552 –1562.

Bird, R., McCrae, M. and Beggs, J. (1987) Are gamblers really risk-takers?Australian Economic Papers, December, 237 – 253.

Bird, R. and McCrae, M. (1994) Efficiency of racetrack betting markets:Australian evidence. In: Efficiency of Racetrack Betting Markets (1994),eds Hausch, D.B., Lo, S.Y. and Ziemba, W.T., pp. 575 – 582, AcademicPress, London.

Blackburn, P. and Peirson, J. (1995) Betting at British racecourses: an analysisof semi-strong form efficiency between bookmaker and tote odds,Studies in Economics, University of Kent, 95/4.

Bolton, R.N. and Chapman, R.G. (1986), Searching for positive returns atthe track: a multinomial logit model for handicapping horse races,Management Science, 32, 1040 – 1059.

235

BIBLIOGRAPHY

Brown, W.O. and Sauer, R.D. (1993), Does the basketball market believe inthe ‘hot hand’? Comment, American Economic Review, 83 (Dec.), 1377– 86.

Bruce, A.C. and Johnson, J.E.V. (1992) Towards an explanation of betting asa leisure pursuit, Leisure Studies, 11, 201 – 218.

Busche, K. (1994) Efficient market results in an Asian setting. In: Efficiencyof Racetrack Betting Markets, eds Hausch, D.B., Lo, S.Y. and Ziemba,W.T., pp. 615 – 616, Academic Press, London.

Busche, K. and Hall, C. D. (1988) An exception to the risk preferenceanomaly, Journal of Business, 61, 337 – 346.

Cain, M., Law, D. and Peel, D.A. (1996) Insider Trading in the Greyhound BettingMarket, Salford Papers in Gambling Studies, Vol. 1, University ofSalford, Salford.

Cain, M., Law, D. and Peel, D. (2000) The favourite – longshot bias andmarket efficiency in UK football betting, Scottish Journal of PoliticalEconomy, 47 (1), 25 – 36.

Cain, M., Law, D. and Peel, D., Bias and Insider Trading in a Variety ofBetting Markets, Bulletin of Economic Research (forhtcoming).

Camerer, C. (1989), Does the basketball market believe in the ‘hot hand’?American Economic Review, 79, 1257 – 61.

Canfield, B., Fauman, B.C. and Ziemba, W.T. (1987) Efficient marketadjustment of odds prices to reflect track biases, Management Science,33, 1428 – 1439.

Chadha, S. and Quandt, R.E. (1996), Betting bias and market equilibriumin racetrack betting, Applied Financial Economics, 6 (3), 287 – 292.

Chapman, R.G. (1994) Still searching for positive returns at the track:empirical results from 2000 Hong Kong races. In: Efficiency of RacetrackBetting Markets, eds Hausch, D.B., Lo, S.Y. and Ziemba, W.T., pp. 173– 181, Academic Press, London.

Clotfelter, C. and Cook, P.J. (1991) Lotteries in the real world, Journal of Riskand Uncertainty, 4, 227 – 232.

Clotfelter, C. and Cook, P.J. (1993) The ‘Gambler’s Fallacy’ in lottery play,Management Science, 39 (12), 1521 – 1525.

Crafts, N.F.R. (1985) Some evidence of insider knowledge in horse racebetting in Britain, Economica, 52, 295 – 304.

Crafts (1994) Winning systems? Some further evidence on insiders and outsiders inBritish horse race betting. In: Efficiency of Racetrack Betting Markets, eds Hausch,D.B., Lo, S.Y. and Ziemba, W.T., pp. 545 – 549. Academic Press, London .

236

BETTING TO WIN

De Bondt, W. and Thaler, R. (1985), Does the stock market overreact?, Journalof Finance, 40, (3), 793 – 805.

Dowie, J. (1976) On the efficiency and equity of betting markets, Economica,43, (170), May, 139 – 150.

Dowie, J. (1982) Why Spain should win the World Cup, New Scientist, June10, 693 – 695.

Figgis, E.L. (1951) Focus on Gambling, Barker, LondonFingleton, J. and Waldron, P. (1996) Optimal determination of bookmakers’

betting odds: theory and tests, Trinity Economic Papers, TechnicalPaper No. 9, December, Trinity College, Dublin.

Gabriel, P.E. and Marsden, J.R. (1990) An examination of market efficiencyin British racetrack betting, Journal of Political Economy, 98, 874 – 885.

Gabriel, P.E. and Marsden, J.R. (1991) An examination of market efficiencyin British racetrack betting: errata and corrections, Journal of PoliticalEconomy, 99, 657 – 659.

Gandar, J.M., Zuber, R.A., O’Brien, T and Russo, B. (1988), Testing rationalityin the point spread betting market, Journal of Finance, 43 (Sept.), 995– 1008.

Gandar, J.M., Dare, W.H., Brown, C.R. and Zuber, R.A. (1998), Informedtraders and price variations in the betting market for professionalbasketball games, Journal of Finance, 53 (Feb.), 385 – 401.

Gilovich, T., Vallone, R., & Tversky, A. (1985). The hot hand in basketball:On the misperception of random sequences. Cognitive Psychology, 17,295 – 314

Golec, J. and Tamarkin, M. (1995) Do bettors prefer longshots because theyare risk-lovers, or are they just over-confident? Journal of Risk andUncertainty, 11, 51 – 64.

Golec, J. and Tamarkin, M. (1998), Bettors love skewness, not risk, at the horsetrack, Journal of Political Economy, 106, 1 (March), 205 – 25.

Gray, P.K. and Gray, S.F. (1997), Testing market efficiency. Evidence fromthe NFL sports betting market, Journal of Finance, 52 (Sept.), 1725 –37.

Griffith, R.M. (1949) Odds adjustment by American horse-race bettors,American Journal of Psychology, 62, 290 – 294.

Haigh, J. (1999), (Performance) index betting and fixed odds, The Statistician,48, 3, 425 – 434.

Haigh, J. (1999), Taking Chances: Winning with Probability Oxford UniversityPress, Oxford

237

BIBLIOGRAPHY

Hamid, S.S., Prakash, A.J. and Smyser, M.W. (1996) Marginal risk aversionand preferences in a betting market, Applied Economics, 28, 371 –376.

Hausch, D.B., Lo, S.Y. and Ziemba, W.T. (1994) Efficiency of Racetrack BettingMarkets, Academic Press, London

Hausch, D.B., Ziemba, W.T. and Rubinstein, M. (1981) Efficiency of themarket for racetrack betting, Management Science, 27, 1435 – 1452.

Henery, R.J. (1985) On the average probability of losing bets on horses withgiven starting price odds, Journal of the Royal Statistical Society, 148(4), 342 – 349.

Hurley, W. and McDonough, L. (1995), A note on the Hayek hypothesisand the favorite – longshot bias in parimutuel betting, AmericanEconomic Review, 85, 949 – 55.

Isaacs, R. and Jackson, D.A. (1994), Focus on sport: index betting on sports,The Statistician, 43, 2, 309 – 315.

Johnson, J.E.V. and Bruce, A.C. (1993) Gluck’s second law: an Empiricalinvestigation of horserace betting in early and late races, PsychologicalReports, 72, 1251 – 1258.

Kahneman, D. and Tversky, A. (1973) On the psychology of prediction,Psychological Review, 80, 237 – 251.

Kahneman, D. and Tversky, A. (1982) Intuitive prediction: biases andcorrective procedures. In: Judgment Under Uncertainty: Heuristics andBiases, eds Kahneman, D., Slovic, P. and Tversky, A., pp. 414 – 421,.Cambridge University Press, Cambridge.

Kahneman, D. and Tversky (1984) Choices, values and frames, AmericanPsychologist, 39, 341 – 350.

Kelly, J. (1956) A new interpretation of information rate, Bell SystemTechnology Journal, 35, 917 – 926.

Kopelman, R.E. and Minkin, B.L. (1991) Toward a psychology of parimutuelbehaviour: test of Gluck’s laws, Psychological Reports, 68, 701 – 702.

Krueger, T.M. and Kennedy, W.F. (1990) An Examination of the Super Bowlstock market predictor, The Journal of Finance 45 , 691 – 698.

Ladbrokes Pocket Companion, 1990 Flat Edition, Aesculus Press, Oswestry.Lo, V.S.Y. and Busche, K. (1994) How accurately do bettors bet in doubles?

In: Efficiency of Racetrack Betting Markets (1994), eds Hausch, D.B., Lo,S.Y. and Ziemba, W.T., pp. 615 – 616. Academic Press, London.

Losey, L. and Talbott, J.C. (1980) Back on track with the efficient marketshypothesis, Journal of Finance, 35, 1039 – 1043.

238

BETTING TO WIN

Metzger, M.A. (1985) Biases in betting: an application to laboratory findingsPsychological Reports, 56, 883 – 888.

McGlothlin, W.H. (1956) Stability of choices among uncertain alternatives,American Journal of Psychology, 69, 604 – 619.

Paton, D. and Vaughan Williams, L. (2001) Monopoly rents and price fixingin betting markets, Review of Industrial Organization, November, 19,(3), 265 – 278.

Paton, D, Siegel, D. and Vaughan Williams, L. (2002), A policy response tothe E-commerce revolution: the case of betting taxation in the UK,Economic Journal, 112, June, 296 – 314.

Paton, D. and Vaughan Williams, L. (1998) Do betting costs explain bettingbiases? Applied Economics Letters, 5, 333 – 335.

Paton, D., Vaughan Williams, L. and Fraser, S. (1999) Regulating insider tradingin betting markets, Bulletin of Economic Research, 51, (3), 237 – 241.

Paton, D. and Vaughan Williams, L. (2002) Identifying irregularities in afinancial market, Applied Financial Economics, (12, 633 – 637).

Paton, D, Siegel, D. and Vaughan Williams, L. (2001), 'Gambling Taxation:A Comment', Australian Economic Review, 34(4): 427-440.

Piron, R. and Smith, L.R. (1995) Testing risklove in an experimental racetrack,Journal of Economic Behavior and Organization, 27, 465 – 474.

Pope, P.F. and Peel, D. (1989) Information, prices and efficiency in a fixed-odds betting market, Economica, 56, 323 – 341.

Preston, M.G. and Baratta, P. (1948) An experimental study of the auctionvalue of an uncertain outcome, American Journal of Psychology, 61,183 – 193.

Quandt, R.E. (1986), Betting and equilibrium, Quarterly Journal of Economics,101, 201 – 207.

Sauer R.D. (1998), The economics of wagering markets, Journal of EconomicLiterature, 36, 2021 – 2064.

Schnytzer, A. and Shilony, Y. (1995) Inside information in a betting market,Economic Journal, 105, 963 – 971.

Shin, H.S. (1991) Optimal betting odds against insider traders, EconomicJournal, 101, 1179 – 1185.

Shin, H.S. (1992) Prices of state-contingent claims with insider traders, andthe favourite – longshot bias, Economic Journal, 102, 426 – 435.

Shin, H.S. (1993) Measuring the incidence of insider trading in a marketfor state-contingent claims, Economic Journal, 103, 1141 – 1153.

Slovic , P. (1972) Psychological s tudy of human judgement:

239

BIBLIOGRAPHY

implications for investment decision making, Journal of Finance,27, 779 – 799.

Snyder, W.W. (1978) Horse racing: testing the efficient markets model, Journalof Finance, 33, 1109 – 1118.

Snyder, W.W. (1978) Decision-making with risk and uncertainty: the caseof horse racing, American Journal of Psychology, 91 (2), June, 201 – 209.

Stock, D. (1990) Winner and loser anomalies in the German stock market,Journal of Institutional and Theoretical Economics, 146 (3), 518 – 529.

Swidler, S. and Shaw, R. (1995) Racetrack wagering and the uninformedbettor: a study of market efficiency, Quarterly Review of Economics andFinance, 35 (3), 305 – 314.

Terrell, D. (1994) A test of the Gambler’s Fallacy: evidence from pari-mutuelgames, Journal of Risk and Uncertainty, 8, 309 – 317.

Terrell, D. and Farmer, A. (1996) Optimal betting and efficiency inparimutuel betting markets with information costs, Economic Journal,106, 846 – 868.

Thaler, R. and Ziemba, W. (1988) Anomalies – parimutuel betting markets:racetracks and lotteries, Journal of Economic Perspectives, 2, 161 – 174.

Thaler, R. (1985) Mental accounting and consumer choice, Marketing Science,4, 199 – 214.

Tuckwell, R. (1983) The thoroughbred gambling market: efficiency, equityand related issues, Australian Economic Papers, 22, 106 – 108.

Vaughan Williams, L. (1999) Information efficiency in betting markets: asurvey, Bulletin of Economic Research, 51, (1), January, 1 – 30.

Vaughan Williams, L. (1999) Monopoly power, information and wageringmarkets. In: The Current State of Economic Science, ed. S.B. Dahiya,pp.559 – 568. Spellbound Publications, Rohtak.

Vaughan Williams, L. (1999) The costs of insider trading: evidence fromdefined markets, Economic Issues, September, 19 – 30.

Vaughan Williams, L. (2000) Can forecasters forecast successfully? Evidencefrom UK betting markets, Journal of Forecasting, 19, 505 – 513.

Vaughan Williams, L. (2000) Markets and information efficiency: evidencefrom the UK Global Business & Economics Review – Anthology 2000,pp. 24 – 29.

Vaughan Williams, L. (2001), Can bettors win? A perspective on theeconomics of betting, World Economics, January, 2, (1), 31 – 48.

Vaughan Williams, L. (2003), The Economics of Gambling, Routledge,London.

240

BETTING TO WIN

Vaughan Williams, L. and Paton, D. (1996) Risk, return and adverseselection: a study of optimal behaviour under asymmetricinformation, Rivista di Politica Economica, 11 – 12, 63 – 81.

Vaughan Williams, L. and Paton, D. (1997) Why is there a favourite –longshot bias in British racetrack betting markets? Economic Journal,107, January, 150 – 158.

Vaughan Williams, L. and Paton, D. (1997) Does information efficiencyrequire a perception of information inefficiency? Applied EconomicsLetters, 4, 615 – 617.

Vaughan Williams, L. and Paton, D. (1998) Why are some favourite –longshot biases positive and others negative? Applied Economics, 30(11), 1505 – 1510.

Vaughan Williams, L. and Sauer, R.D., Information Efficiency in GamblingMarkets (draft title), Cambridge University Press, Cambridge,forthcoming

Vergin, R.C. (1977) An investigation of decision rules for thoroughbredracehorse wagering, Interfaces, 8, (1), November, 34 – 45.

Wells Fargo-Nikko Investment Advisors, Global Currents, March 1995, p.1.Cited in notes to Chapter 15 of: Against the Gods: The Remarkable Storyof Risk, Bernstein, P.L., (1996), John Wiley & Sons, Inc., New York

Woodland, L.M. and Woodland, B. M. (1994) Market efficiency and thefavourite – longshot bias: the baseball betting market, Journal ofFinance, 49, (1), 269 – 279.

Zuber, R.A., Gandar, J.M. and Bowers, B.D. (1985) Beating the spread: testingthe efficiency of the gambling market for National Football Leaguegames, Journal of Political Economy, August, 800 – 806.

Documents

Betting